dennisthiessen/claude-resume-kit
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Compare commits

base: dennisthiessen/claude-resume-kit:94913a004a12ffdcce90cad29ccc9f4cc7760bdb
Branches Tags
dennisthiessen/claude-resume-kit:main
...
compare: dennisthiessen/claude-resume-kit:main
Branches Tags
dennisthiessen/claude-resume-kit:main
33 Commits
94913a004a ... main
Author SHA1 Message Date
dennisthiessenandClaude Opus 5 89af779bf3 docs(sbb): add total-comp break-even worksheet 
The deciding question on 2026-09-09 is a number nobody has, so the
worksheet is a frame to complete before the call rather than a filled-in
answer: Anforderungsniveau K is not public, which is why Q1 exists.

Records the three things that make a Swiss base-to-base comparison lie:
the free GA, Regionalzulage and employer pension share are real money
invisible in a base figure, while a Swisscom bonus is real money SBB
will not match; the GA is a taxable Lohnausweis benefit so its net value
sits below list price; and the GA and the "commute saving" are the same
franken, so counting both flatters SBB by four figures.

K is a capped GAV band with no negotiation lever at the top, so this is
a yes/no against a fixed number, not an opening position.

User's read is that it will not clear. Taking the call regardless: the
questions are worth asking, it is interview practice while NATO JWC,
BIS and Microsoft Principal are still live, and the K figure is the
first real data point on what the entire Bern/Thun local tier pays -
RUAG, Swissgrid, BKW and BFH sit in comparable band structures. Worth
recording whatever the outcome.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JMcHCsTKWvVzqyLChF5ckk
 2026-08-28 17:01:14 +02:00
dennisthiessenandClaude Opus 5 9a188cafef feat(sbb): interview invitation 2026-09-09, write interview brief 
SBB Data Engineer (Job ID 103755): invited 2026-08-28 for Wednesday
9 September 2026, 08:30, 45 min on Teams.

Per the posting's own process this is stage 2 of 4 - "Virtuelles
Kennenlernen mit HR und Fuehrungskraft" - so HR together with Andri
Wienandts, not a pure HR screen. 45 minutes shared between two people,
which is the practical constraint the brief is built around.

Two facts worth recording:
- This is the FIRST conversion of the evidence-first cohort (6
  applications, 2 rejections), and it came from a COLD submit. The
  channel plan called for phoning Wienandts before applying; that never
  happened, so the published warm contact went unused.
- The resume never contained Snowflake, dbt or Power BI - the three
  literals an ATS keyword screen would have keyed on, and a risk the
  session had explicitly recorded as "accepted, not solvable". It did
  not filter him out, so a human read the dossier and the honest
  substitution framing survived first contact. One data point; not
  enough to revise strategy on, but it is evidence against assuming
  honest omissions are fatal at the screen.

Brief covers the four questions that must not go unasked (Anforderungs-
niveau K band, design/architecture scope vs current Staff + Component
Owner level, RAMSI's Java/Spring Boot/Angular share, Kidz Care rate),
the substitution answers for each named-but-non-canonical tool, the
swissTAMP/RIS/MUD context, and the thesis boundary - methods prototype,
no operational data, no accuracy figures, PSO surveyed only.

It also asks him to decide three things before the call rather than on
camera: what he does if K lands at 130-150k, if architecture is set
elsewhere, or if RAMSI turns out to be heavily Angular. Any one can be
a no, and the level question is the same shape as the declined BKW.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JMcHCsTKWvVzqyLChF5ckk
 2026-08-28 16:41:55 +02:00
dennisthiessenandClaude Opus 5 a0f844c143 feat(ruag): submit AI Engineer C5I, and record the Kdo Cy package 
RUAG C5I - AI Engineer C5I (Thun, app ID 18027): SUBMITTED 2026-08-27.
Evidence Fit 74/100, fit class Stretch, title-capability hard gate FAIL
(R1 build/operate the AI-LLM platform and R2 in-house LLM/retrieval are
Adjacent, not Direct) - all knowingly accepted under the user's explicit
Phase 0 override. Consumes the cohort's sole Stretch slot: cohort is now
6/10, Stretch 1/1.

Critique ran twice. Round 1 scored the documents 85/100 and raised three
Tier 1 items; Round 2, after the fixes, scores 93/100 with truth and
provenance 24/25 and zero Tier 1 findings. Evidence Fit did not move and
was not allowed to: PP-3 is a personal project and cannot convert an
Adjacent title-capability into a Direct one.

Documents (.tex is the deliverable; PDFs are gitignored):
- resume 2pp/13 bullets, letter 1pp/295 words, both validators PASS
- headline is now the canonical title "Staff Data, Analytics & AI
  Engineer", which puts the req's own word above the fold
- PP-3 added as a Projekte section - the only current Linux/hardening
  evidence, since BS-6 ended Dec 2022
- the letter names the AI-platform gap in one sentence, then connects
  SW-5 to C5I's stated DevSecOps model
- "governte" -> "governance-konforme"; JD term coverage 18/22 -> 22/22
  claimable, with all six gap terms still correctly absent

Two defects found and fixed that the first critique missed:
- the resume never loaded babel, so a German document was hyphenated
  with English patterns (Hal-bleiterfertigung, Tran-skription). Fixed
  with babel[ngerman] plus a Transkription exception; this also cleared
  an overfull box. Check this on every future German package -
  resume_template.tex likely has the same omission.
- IBM AI Engineering had no primary-source record. Certificate read
  directly (IBM via Coursera, 4 Jun 2020, verify 3ZBZFVAL6A34), recorded
  as entry #8; it also proves PyTorch, now evidence: certification.

Also included: the submitted Kdo Cy DevOps Engineer III package, the
Capgemini omit rule promoted into config.md, BW-1 canonicalized in
experience_bundeswehr.md, and a scout.py comment correcting the
telenorgroup slug from "near-empty" to a claimed board serving stale
phantom listings that DEMO_TITLES would not catch.

Compensation, PSP/project eligibility and role level were never resolved
before sending and are now live screening topics.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JMcHCsTKWvVzqyLChF5ckk
 2026-08-27 23:20:17 +02:00
dennisthiessenandClaude Opus 5 93b0fc5b76 chore(akerbp): record rejection, cold channel and slot stays consumed 
Rejected 2026-08-27, no interview, ~28 days after submission. Recorded
in all four places that track an outcome: cohort state, decision log,
the Active Sessions row and the session file's status block.

The cohort slot is not freed. A rejection consumes an attempt the same
way a submission does, so the Adjacent count stays 2/2 and the cohort
stays 4/10 - unlike the Google FDE III and AWS FDE no-gos, which were
declined before a package existed.

Recorded as rejected_no_interview on the evidence: nothing in the log
shows any contact after submission. Recruiter-vs-hiring-manager
disposition is a tracked cohort variable, so this is one edit away from
being corrected if a screen did happen.

The package was validator PASS with no Tier 1 truth findings and the
follow-up email to Per Olav Marthinsen was never actioned, making this a
pure cold submit. A cold-channel no-interview rejection does not
implicate document quality on its own. That is now 2 of 4 cohort
applications rejected without interview, both cold, which fits the
channel thesis already in the log - and application_strategy.md forbids
reacting to a single rejection, so no positioning or targeting changed.

The Norway lane is explicitly not closed. NATO JWC Stavanger is still
live and strong Norwegian employers remain in scope; both the row and
the decision note say so, because with Aker BP closed the earlier
Equinor note ("lane now reduced to the open Aker BP application") would
otherwise read as an implied closure.

Submission date left inconsistent on purpose: the contemporaneous
records say 2026-07-29, the retrospective ones 2026-07-30. Both now
carry a note pointing at the other rather than one being overwritten.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JMcHCsTKWvVzqyLChF5ckk
 2026-08-27 14:45:09 +02:00
dennisthiessenandClaude Opus 5 420e205e37 fix(kb): settle ECTS B as the 10-35% band, 'top 35%' is correct 
ECTS grade B is the next 25% after the top 10%, so the accurate phrasing
is 'top 35%'. The record's original value was right; the 'top 30%'
considered yesterday was the error, now removed rather than left as an
open question.

The field also now says never to narrow B to a point estimate. B is a
band, and quoting 'top 30%' or 'top 25%' overstates the standing - the
kind of small, checkable claim that costs credibility for no gain.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:56:34 +02:00
dennisthiessenandClaude Opus 5 124f5d4953 feat(kb): record degree final grades and US 4.0 conversions 
A Citadel application form asked for GPA and neither degree's
Abschlussnote existed anywhere in claims.json - only the M.Eng. thesis
grade. The single lead was an old CV extraction whose own provenance
note disclaims it and which already carries a known error, so it was not
safe to convert from. User confirmed both directly: M.Eng. 1.6 and
B.Eng. 2.4 on the German 1.0-4.0 scale.

US 4.0 equivalents recorded via the modified Bavarian formula, which
simplifies to 5.0 - G: 3.4 and 2.6. Both fields carry usage rules,
because a converted grade is easy to misuse. They are for application
forms and credential checks only, never to be presented as a natively
awarded GPA, and the German scale must always be stated alongside -
particularly for the bachelor, where the conversion makes a solid "gut"
read weaker to a US audience than it is.

Also flagged a percentile that does not agree with itself. The relative
grade is ECTS B, but this record said top 35%, the user said top 30%,
and ECTS B is formally the next 25% after the top 10%. Rather than pick
one, the field now says to check the certificate and notes that quoting
"ECTS grade B" alone is always safe. A specific percentile on an
application is exactly the kind of small number that a credential check
can contradict.

New grade_output_rule: neither grade belongs on a resume or in a cover
letter by default. Industry resumes omit grades, and the thesis 1.0 is
the only figure worth surfacing where academic performance is valued.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:53:55 +02:00
dennisthiessenandClaude Opus 5 5cd5f49aa5 chore(citadel): record submission, and the form had no Zurich option 
Submitted 2026-08-26. Cohort entry added (Adjacent 2/2, cohort now
4/10, channel weak x3) and decision logged.

The finding that matters is not the submission. The application form's
preferred-work-location selector offered no Zurich, despite the JD body
reading "in Miami, Zurich or New York", the posting header listing
Zurich, and the site's own Zurich location filter returning this exact
req when checked on 2026-08-25. The careers site footer reads Citadel
Enterprise Americas LLC, so the likeliest explanation is a US-entity
apply flow with European seats routed elsewhere, or a Zurich seat that
is closed while the multi-location posting stands.

This partly supersedes the working-model risk that blocked Phase 2. If
no Zurich seat is reachable through that form, whether Zurich runs
hybrid or five-day onsite is moot, and the application may be sitting
against a US requisition - a hard no under the standing no-relocation
constraint.

Carry-forward rule recorded in both CLAUDE.md and the session file:
check the application form's location selector before investing in a
package. Two independent signals - the posting's stated locations and
the employer's own location filter - both proved unreliable as evidence
that a seat is actually reachable. Phase 0 treated the filter result as
confirmation that the req was live in Zurich; that inference was weaker
than it looked.

Package itself is finished and reusable: both documents cleared the
validator with zero warnings and the Tier 1 fixes lifted JD term
coverage from 51% to 68%. Nothing about the documents is the weak point.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:49:08 +02:00
dennisthiessenandClaude Opus 5 0c34bd0347 chore(citadel): finalize package, hold submission records 
Finalization per shared_ops.md. Both submitted documents re-validated:
resume and cover letter each PASS with zero warnings. Final PDFs copied
to Dennis_Thiessen_Resume.pdf and Dennis_Thiessen_Cover_Letter.pdf.

Steps 5 and 6 - submission date and cohort entry - are deliberately not
done. The user directed "apply as is" but has not confirmed the
application was actually sent, and stamping today's date on a submission
that may happen tomorrow corrupts the log that every cohort and
disposition-timing judgement is read from. Both commands are written out
in the session file, ready to run on confirmation. This is the same
discipline applied to the SBB Wienandts call: record what happened, not
what was intended.

The critique stays STALE by choice rather than being re-run: it scored
the pre-Tier-1 resume and predates the cover letter. The estimated ~91
post-edit Document Quality is flagged in both the session file and here
as an estimate that must never be quoted as a score.

The working-model question is unresolved and now travels with the
application rather than blocking it, which is the user's call and a
reasonable one now that the package exists: clarifying first only pays
off in the branch where a cold application gets a call. It must be asked
at first recruiter contact. If Zurich is 5-day onsite this is a decline
on the Bern constraint no matter how the process goes.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:34:28 +02:00
dennisthiessenandClaude Opus 5 57c5a388a9 docs(citadel): record cover letter in session file 
The session update in f8b4cd3 silently failed - the heredoc ran from a
stale working directory left by an earlier cd, so the commit landed
without it. Output Files and Cover Letter status now recorded.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:31:53 +02:00
dennisthiessenandClaude Opus 5 f8b4cd3473 feat(citadel): cover letter, and fix SW-1 scope traps in two more bundles 
Cover letter: 1 page, 270 words, 3 paragraphs, validator PASS. Carries
the one thing the resume structurally cannot - why a Swisscom data
engineer credibly wants a research-platform seat at a market maker,
which against an ex-FAANG field is the main differentiator.

Deliberately uses no external hooks. cl_reference.md says to omit an
unnecessary hook rather than spend words proving company familiarity, so
the only hook is the JD's own language, scraped verbatim and first-party.
No named executives: nothing to verify, and nothing that reads as
name-dropping. The letter also states the R1 gap plainly rather than
hiding it, because a technical reviewer will find it in the first
question anyway and owning it is stronger than being caught by it.
PP-1 appears with its personal-project label and an explicit "I make no
claims for its results".

Separately, loading the bundles for this letter surfaced two live scope
traps. The 2026-08-21 correction that fixed the SW-1 violation in
bundle_data_engineer.md ended with "check the other four bundles for the
same pattern". That sweep was never done, and two of them carried it:

  bundle_data_platform.md - the S5 cover-letter hook read "migrating
  Swisscom's legacy ETL stack to a cloud-native AWS platform", and its
  narrative thread said "migrating an entire ETL infrastructure". Both
  pair a full-ownership framing with a company-scale object, which
  claims.json SW-1 forbids outright.

  bundle_ml_ai_engineer.md - the SW-1 reframing row read "Built
  cloud-native data infrastructure on AWS ... the scalable data layer",
  which is both a full-ownership verb on an org-scale object and an
  unverified scale claim.

Both rewritten to the scoped form with inline scope warnings, matching
the data_engineer fix. bundle_analytics_engineer.md and
bundle_semiconductor.md were checked and are clean.

These were loaded traps: any future cover letter built from either
bundle would have started from a sentence claims.json forbids.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:31:39 +02:00
dennisthiessenandClaude Opus 5 c94edbce44 fix(citadel): apply Tier 1 relevance fixes, JD coverage 51% -> 68% 
All three Tier 1 items from the critique, all honestly available, no
claim altered.

The R2 vocabulary now appears where R2 is actually evidenced: the SW-7
bullet carries ingestion and lifecycle, and SW-1 carries transformation
and storage. R2 was the single core responsibility scored Direct, and
the document had been proving it in substance while missing it in the
JD's own words. SW-2 now names the Oracle/Kafka/Python/Teradata estate a
distributed system he operates - not one he designed, which stays a gap.
The second skills line is relabelled "Distributed systems" and names
distributed databases explicitly, closing the fifth preferred-technology
hit that Phase 0 counted as Direct but the resume never actually wrote.

Recorded what was deliberately not added, because raising keyword
coverage is exactly where a later run would be tempted to overreach:
"scalable" is an unverified scale claim, and SDK, high-throughput,
data-intensive, simulation, model development and self-service have no
canonical evidence. Verified still absent after the edit.

Scope kept to what was approved. The user said "tier 1 fixes", so the
Tier 2 headline retitle - "Production Data Platforms on AWS", the line
most likely to draw the opening challenge from a technical reviewer -
was left in place for them to decide.

Validator PASS with zero warnings: the 58% cadence warning cleared as a
side effect, because restructuring two bullets with em-dashes broke the
comma-list pattern the checker keys on. Still exactly 2 pages, zero
overfull/underfull boxes, PDF re-inspected.

Critique marked STALE - it scored the pre-edit document.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:17:47 +02:00
dennisthiessenandClaude Opus 5 7953884810 feat(citadel): critique the resume - Document Quality 86/100 
No Tier 1 truth findings, validator PASS, and the claim audit clears
every material claim including the PP-1 block, which carries its
personal-project label three times and no performance figure.

The one real weakness is that the document under-uses the JD's own
language for the responsibility it matches best. R2 - ingestion,
transformation, storage and lifecycle management of large datasets - is
the single core responsibility scored Direct and is close to a plain
description of SW-1 and SW-7, yet none of those four words appears in
the resume body. Two related misses: "distributed systems" never appears
as a phrase even though Q4 is a required qualification scored Direct, so
the nearest exact match sits on an eight-year-old Vizrt bullet rather
than current work; and "distributed databases" was called a Direct
preferred hit in Phase 0 but is only ever implied through product names.
Three Tier 1 fixes, all honestly available, worth roughly +4.

Recorded what must NOT be added, since the obvious way to raise keyword
coverage is the wrong one: "scalable" is an unverified scale claim that
ai_fingerprint_rules forbids, and SDK, high-throughput, data-intensive,
simulation, model development and self-service have zero canonical
evidence. Their absence is correct, not a defect.

Flagged the headline "Production Data Platforms on AWS" as Tier 2. It
carries no ownership verb so it is not a scope violation, but it is the
most prominent line in the document and leans toward exactly the
platform framing R1 is a gap on - the likeliest opening challenge from a
technical reviewer.

Scored against critique_framework.md section 9. The SKILL.md
8-dimension table with "Publications 10%" is a stale CV-era scheme and
was not used, the same conflict class as its "all bullets 2L" line.

The verdict does not change the decision. Document quality is not the
constraint here: the working model is still unresolved and the channel
is still cold. 86 is not submit-ready.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 09:06:18 +02:00
dennisthiessenandClaude Opus 5 89f9af5d89 feat(citadel): 2-page resume for Citadel Securities Platform Engineer 
13 bullets, validator PASS, compiles to exactly 2 pages, rendered PDF
inspected clean - no clipping, orphans, header wrapping or bad breaks.

Built at the user's direction with the working-model question still
open. That block stands: if the Zurich seat is 5-day onsite, this
package does not get submitted. Recorded in the session file rather than
quietly dropped.

Three Phase 2 decisions the user left open, taken and marked reversible.
VZ-1 is in - it is the only canonical evidence pairing C++ with a
distributed backend and this JD names both, hedged verb preserved. A
Personal Project section carries PP-1, labelled three separate ways and
with no performance figure of any kind; PP-2 appears as supporting
coursework, never as a credential. Generali is kept against the Phase 1
"omit" recommendation, because dropping it opened an unexplained
May 2015 to Jun 2017 employment gap that costs more than the weak bullet.

Cadence sits at 7/12 (58%), above the section 7a guide of ~50%, down
from 75% after reshaping two bullets. The rest are legitimate technology
enumerations - "Oracle, Kafka, Python and Teradata", "Elasticsearch,
Logstash, Kibana and Kafka". The checker added yesterday cannot tell a
rhetorical rule-of-three from a list of tools actually used, which is a
real limitation of the pattern rather than a defect in this document.
Cutting them would delete accurate ATS-relevant names, which section 7a
itself forbids. Left deliberately.

Also walked into the trap the SBB .tex header warns about: the first
draft named the excluded non-canonical tools in a comment explaining why
they were excluded, and the validator errored on all three because it
scans raw file text. Comment rewritten without naming them.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 08:42:46 +02:00
dennisthiessenandClaude Opus 5 0eda29327a fix(citadel): working model unresolved - block Phase 2, 74 -> 71 
The user asked whether Citadel has a hybrid model. The JD is silent:
grepping the verbatim posting for hybrid, remote, onsite, in-office and
days per week returns zero matches. Phase 0 scored practical constraints
5/5 on an unexamined assumption that Zurich was workable. Withdrawn.

Ken Griffin is on record that Citadel returned to the office five days a
week, early and deliberately, and calls it his most important leadership
decision of the past seven years - framed as core apprenticeship culture
rather than a policy setting. The only contrary evidence is an anonymous
2022 Fishbowl post about Operations claiming three days minimum: four
years stale, wrong division, contradicted by the founder on record. No
official Citadel Securities Zurich policy statement could be found.

If five-day onsite holds for Zurich, it means roughly 2 to 2.5 hours of
commuting per day from Bern, or relocating - which breaks both the
Bern-based, 2-3 day hybrid bar in user_comp_bar and the "keep a Swiss
home base" constraint in user_international_mobility.

Practical constraints 5/5 -> 2/5 (language and authorization stay clean:
English working language, EU citizen, Swiss B permit). Evidence Fit
74 -> 71, still Adjacent but mid-band rather than a point below Core.

More consequential than the number: critique_framework.md treats an
unresolved location or relocation constraint as a NO-GO trigger, so the
qualifications gate passes while the practical gate does not. Phase 2 is
blocked pending an answer.

This has to be asked, not researched - the same class of question as the
SBB Anforderungsniveau K band. Recording the sequencing explicitly,
because SBB was submitted with its equivalent question still open and
that question is now a live screening topic instead of a settled fact.
Repeating it here should be a deliberate choice, not an oversight.

The Phase 1 bullet plan is written and unaffected: 11 recommended, 14
with options, budget PASS. Two items still open - the Vizrt VZ-1
C++/distributed bullet the user previously skipped on Snowflake, and
adding a Projects section for PP-1.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-26 08:30:22 +02:00
dennisthiessenandClaude Opus 5 ec5be5f2ed feat(citadel): Phase 0 for Citadel Securities Platform Engineer, Zurich 
Evidence Fit 74/100, Adjacent at the top of the band, hard gate PASS
with one flagged risk. Phase 1 not started - user asked for Phase 0 only.

The stack alignment is the best in the log: zero Gaps across all seven
minimum qualifications, and all five preferred technologies (Kafka,
Kubernetes, Spark, Airflow, distributed databases) are Direct and
production-current.

What holds it at 74 is that the title-defining function splits. R2 -
ingestion, transformation, storage and lifecycle management of large
datasets - is Direct and describes SW-1/SW-7 almost word for word. R1 -
design and build the distributed research platform itself - is Adjacent,
the same authoring-vs-building shape as Aker BP, and it cannot be
written around: Scope Discipline forbids pairing a full-ownership verb
with an org-scale object, so it can only be honestly bridged.

Recorded the classification conflict rather than hiding it.
application_strategy.md says a defining responsibility that is only
Adjacent makes the role Stretch; that strict reading applies to R1. It
is filed as Adjacent because R2 is equally title-defining and is Direct,
and because Aker BP set that precedent at 79. Both cohort slots are free
either way, so the label changes nothing operationally.

Corrections the user was right about, now recorded so a later run does
not repeat them: Go is NOT a gap, because Q3 is disjunctive and Python
satisfies it outright; and a production Kafka/K8s/Spark/Airflow estate
IS a distributed system, so Q4 is Direct rather than a stretch.

Two traps written down explicitly. The $175,000-350,000 base range is a
US pay-transparency disclosure under NY law and says nothing about the
Zurich figure - it must be asked, not assumed. And the sibling reqs that
web search surfaces as closer title matches (Research Platform - Data
Platform Engineer, Research Platform Infrastructure Engineer) are dead:
404 and Cloudflare-blocked, absent from the live board. Those Built In
listings are stale.

JD saved verbatim via Playwright; WebFetch returns 403 on this host.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-25 23:24:02 +02:00
dennisthiessenandClaude Opus 5 b97dec978a feat(kb): add trading-domain evidence as PP-1 and PP-2 
claims.json had zero hits for trading, finance or quant, so a real
capability could not legally reach any document: the file is the
highest-authority source and nothing outside it is claimable. Surfaced
while assessing the Citadel Securities Platform Engineer req (Zurich),
where domain fluency is exactly what was missing.

PP-1 is the self-built investing/signal platform: US-equity price,
fundamental and LLM-assisted sentiment ingestion, a long-only
cross-sectional residual 12-1 momentum book with ATR stop/trail and max
15 concurrent names, scheduled scan and backtest pipelines, web
dashboard and Telegram alerts. PP-2 is the Udacity AI for Trading
nanodegree. Two allowed-with-context skills reference both.

Written defensively, because this is the kind of entry a later run
inflates. Each carries an explicit forbidden list: no PnL, Sharpe,
return or backtest figure may EVER be quoted (none is verified); PP-1 is
single-user and self-hosted, so never distributed, scalable, production
or multi-user; neither implies professional quantitative, trading or
finance experience; the nanodegree is never an academic or quant
credential. Legitimate use is self-directed domain fluency plus an
end-to-end ingestion/backtest/alerting build outside work.

PP-1 is also the one claim where full-ownership verbs are correct - it
is genuinely solo, unlike the employer work that Scope Discipline
governs. Stated in the scope so the two rules do not collide later.

JSON valid, validator PASS, 9/9 tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-25 23:18:18 +02:00
dennisthiessenandClaude Opus 5 92a81a7dcb feat(resume): check bullet cadence variety, warn-only 
An audit of all 18 packages in output/ (368 bullets) found one rhythm
running through every document: 45% of bullets used the same "X, Y and
Z" triple, and the SBB package reached 85% - 11 of 13 bullets, every
Swisscom and Bosch line - plus two adjacent bullets both opening
"Build and...".

This is a style finding, not a truth finding. Every bullet was accurate.
The corpus is clean on the axes that actually signal generated text: no
AI vocabulary (0 hits for leverage/spearheaded/robust/passionate and 56
others across 35 documents), prose em-dashes at 0.08/bullet, and PDF
metadata carrying nothing but MiKTeX pdfTeX with empty Author/Title
(0 AI tokens and 0 generator-term leaks across 69 PDFs). What is left is
cadence: ten bullets sharing one three-beat rhythm read as machine-made
even when nothing in them is false.

Guarded deliberately so it cannot do harm. cadence_checks() emits WARN
and never ERROR, the critique deduction caps at 1 point, and both the
reference and the docstring state that no claim, scope or hedged verb
may be bent to satisfy rhythm. An anti-monotony rule with teeth would be
worse than the problem - it would pressure a future run into loosening a
scoped claim to vary a sentence.

Thresholds are calibrated on the corpus, not guessed. Position length
spreads are bimodal (1-9 words, then 15-17), so the check flags a spread
of <=4, the tight tail at ~28% of positions; the first draft used <=6
and flagged the median. Also fixed the opening-verb extractor, which
read \textbf{Owned ...} as the word "textbf" and produced six false
positives on one document. Per-package warnings now run 0-3.

Docs: resume_reference.md 7a + verification step, critical_rules.md 7a,
critique_framework.md mechanics row, CLAUDE.md corrections log.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-25 09:38:19 +02:00
dennisthiessenandClaude Opus 5 af7189a93d chore(scout): version the application cohort tracker 
application_cohort.json was caught by the blanket job_scout/state/*
ignore, so the cohort slot record (core/adjacent/stretch mix and every
application's fit class, evidence fit, hard gate and outcome) existed
only on one machine. decisions.json escaped the same rule only because
it had been tracked before the rule was added - an accident of history,
not a decision.

Both files are durable application history and belong in the repo. The
churny scan state (seen_jobs.json, last_scrape.json) stays ignored, as
does reports/ - all of it is regenerable by re-running the scout.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-25 09:08:14 +02:00
dennisthiessenandClaude Opus 5 e53abd4ff6 chore(log): submit SBB, close Snowflake, correct Norway lane framing 
Application log updates from the 2026-08-24 scout run and the days after.

SBB Data Engineer (Asset Management, Bern, job 103755): SUBMITTED
2026-08-25. Consumes Core slot 2/7 in evidence-first-2026-01. Submitted
with the Anforderungsniveau K comp question still unresolved, against
the channel plan and the critique's closing advice, so the K band, the
design-vs-build-and-run scope question and the Kidz Care rate are now
screening-call topics rather than pre-cleared facts. Next action changed
from "call then submit" to "prep an interview brief". The stale pointer
to an /edit-resume Tier 1 fix is removed - that fix landed 2026-08-21.

Snowflake Sr SWE Enterprise (Observe): CLOSED - NO RESPONSE at 80 days
silent, past the longest real disposition in the log (Equinor, 79).
Recorded as closed_no_response, deliberately NOT as a rejection: none
was ever received, so it must not enter rejection statistics or read as
a document-quality signal. Pre-dates the cohort, so no slot is freed.
An ~86/100 package - the highest in the log - drawing no reply at all is
the strongest evidence yet that the constraint is channel, not documents.
Unblocks the Observe Metrics Platform req, which was paused only to
avoid stacking a second cold application on the same org.

Norway lane: corrected a false "closed" framing that had propagated into
the Equinor row and was used to dismiss Norwegian rows in the scout
readout. The lane is OPEN but selective. The gates are norsk working
language and NO security clearance - NOT compensation: the 180k bar is
CH-only and explicitly does not apply to Norway. Equinor, Telenor and
NATO JWC stay in the scout.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MHtzyTKBcg6BWhD5qFegtK
 2026-08-25 09:07:07 +02:00
dennisthiessenandClaude Opus 5 73336cc4f9 feat(sbb): critique the package and apply its Tier 1 fix 
Critique per critique_framework.md: hard gate PASS, Evidence Fit 79/100
(Core, unchanged — a document cannot move candidate-role fit), Document
Quality 90/100, Channel Weak but uniquely upgradeable. Claim audit clean
across all 13 bullets; no Tier 1 truth findings.

The skill's own spec asks for a single 8-dimension score including a
Publications weight, which contradicts critique_framework.md and CLAUDE.md
("never collapse them into one optimistic score"). Followed CLAUDE.md and
recorded the conflict — the skill definition looks stale relative to the
July rewrite.

Tier 1 finding and fix: the JD names Abfrageleistungsoptimierung in a
required line and the document had zero occurrences of "query", despite
canonical support sitting in BS-2's 3L variant. BS-2 now reads "tuning
query performance for analysis teams working with…" and the Pipelines
skills line gains "query performance tuning". Also surfaced "agile" in
SW-3, whose canonical 2L variant records "in an agile DevOps team" and
whose absence left the JD's second required line unaddressed.

Deliberately NOT applied: the API term. claims.json SW-7 records only
"onboard source systems" and SW-2's ingestion is Oracle to Kafka, so
adding it would be vocabulary substitution rather than evidence. Left
open pending user confirmation.

Re-verified after the edit: validator PASS, forbidden-token scan PASS,
compile PASS at 2 pages, both pages re-rendered and inspected. Longest
bullet 34 words, still under the flag threshold. Document Quality 90 to
92 on relevance/terminology.

Still not submitted — held pending the Anforderungsniveau K comp answer.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WpMtWFtYkGkjSMShuTgXiB
 2026-08-21 17:09:27 +02:00
dennisthiessenandClaude Opus 5 a64661e42a feat(sbb): 2-page English resume for SBB Data Engineer, Bern 
13 bullets, 6 skills lines, International Tech profile in English with an
explicit Languages line naming German as mother tongue (user instruction).
Framing leads on operate-and-own — Component Ownership with on-call and
governance answers the JD's "Betrieb" and "Verantwortung für Qualität und
Nachhaltigkeit" — and carries the Bosch industrial sensor-data thread
(defect records, wafer inspection images, PCM parameters) against the
role's vorausschauende-Wartung purpose. GenAI/LLM material dropped.

Two deviations from the approved 11-bullet plan, both deliberate:
Fraunhofer keeps one combined bullet rather than zero, because dropping the
role opens an unexplained 21-month gap between Vizrt and Bosch; and BS-1 is
added as a fifth Bosch bullet, HIGH priority in the bundle and directly
relevant to an employer running a 24/7 network.

Kept the guarded page break after testing the natural flow: without it page 1
fills, but two Bosch bullets land on page 2 with no visible employer header.
Hierarchy beats page-fill, and page-fill quotas were removed on 2026-07-27.

Gates: canonical validator PASS (0 warnings), forbidden-token scan PASS (none
of the seven non-canonical JD tools present), scope-discipline scan PASS,
compile PASS at exactly 2 pages, both pages visually inspected. SW-1 and SW-7
bullets also read by eye against their claims.json forbidden lists.

Not submitted — held pending the Anforderungsniveau K comp answer.

Also fixes a KB defect found during generation: bundle_data_engineer.md §S5's
cover-letter hook read "leading the migration of our legacy stack", a
full-ownership verb on a company-scale object that contradicts claims.json
SW-1 and its own narrative thread. Rewritten, warned inline, and recorded in
the CLAUDE.md corrections log. The other four bundles are unchecked.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WpMtWFtYkGkjSMShuTgXiB
 2026-08-21 16:56:48 +02:00
dennisthiessenandClaude Opus 5 787a1d937e feat(sbb): Phase 0 for SBB Data Engineer, Asset Management, Bern 
Job ID 103755, posted 2026-08-20, surfaced by the 2026-08-21 scout run.
Evidence Fit 79/100 (Core, lower end), hard gate PASS. The title-defining
capability — build and operate cloud data pipelines and governed data
products — is Direct and current, unlike the AWS FDE NO-GO.

Best practical fit in the log: Bern-based, German-language, no relocation,
EU citizen plus B permit. Real strength cluster for the role's vorausschauende
Wartung purpose: Bosch fab sensor/process data, ELK/Kafka anomaly detection,
containerized ML inference, plus the condition-monitoring thesis.

Two open risks, both level and comp rather than fit: Anforderungsniveau K is
a capped GAV band whose figures are not public and must be asked, and the seat
reads lateral or below current Staff plus Component Owner scope — the same
shape declined at BKW. Level and ownership scored 9/15 for that reason.

Seven named tools are non-canonical (Snowflake, dbt, Argo Workflows, Helm,
Power BI, Spring Boot, Angular); survivable only because the JD prefixes the
stack list with "z. B.". Recorded as an accepted ATS risk, with a Phase 2
gate to keep all seven out of the generated document.

Cover letter: NO — SBB explicitly waives it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WpMtWFtYkGkjSMShuTgXiB
 2026-08-21 16:33:39 +02:00
dennisthiessenandClaude Opus 5 8a30c68c95 fix(scout): key decision lookups by stable job ID, not raw URL 
Boards rewrite their own URLs between runs, so a decision recorded under one
URL was lost the next time the same req appeared. Amazon truncates the job
slug to a fixed width and that width changed, which silently resurfaced the
already-skipped Principal Delivery Consultant req at score 7 in the 08-21 run.

Adds _decision_key(): "<host>#<req-id>" via per-board ID patterns, falling
back to a normalized path. Host-scoping is deliberate — a wrong ID match can
never mark a role decided at a different company. Also collapses genuine
duplicates: Workday's /apply suffix, PostFinance's per-locale URLs, Apple's
?team= and /locationPicker variants, the same NVIDIA req under two locations.

The log file stays URL-keyed (it is hand-edited and the URL is the readable
part); lookups go through index_decisions(). --decide now updates an existing
entry when the job is already logged under a drifted URL instead of adding a
second row.

Verified against all 955 URLs in the decision log plus the 2026-08-21 report:
4 groups collapse, each inspected and correct, no false merges.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WpMtWFtYkGkjSMShuTgXiB
 2026-08-21 16:33:27 +02:00
dennisthiessenandClaude Opus 5 2fadfdb8eb feat(scout): probe four unverified boards — add CERN, NCIA to manual check 
Probed CERN, Swiss Post, NATO NCIA and SIX Group. Two actionable results:

- CERN: SmartRecruiters board "CERN", 66 live Geneva postings with real
  engineering (Deep Learning Developer, ML Engineer NGT, Full-stack SWE, Data
  Storage R&D Engineer). Added — the adapter already existed. Two standing
  caveats recorded in the entry: Geneva is ~1h50 from Bern so a 2-3 day hybrid
  is not viable without relocation, and CERN grades sit below the 180k bar.
- NATO NCIA: nato.int is Cloudflare-blocked to headless browsers. Moved to
  MANUAL_CHECK beside Oracle, with a note that NCIA is a separate pipeline from
  the JWC Stavanger reqs the `nato` adapter returns.

Not added, documented instead:
- Swiss Post: real board is job.post.ch (SuccessFactors CSB), scrapable via
  a[href*='/job/']. The only tech role on page 1 was served from the PostFinance
  sub-board, which is already covered. Paginated, so this is page-1 evidence
  only, not a whole-board verdict.
- SIX Group: SAPUI5 SPA backed by a legacy DWR RPC endpoint
  (careerJobSearchControllerProxy.getInitialJobSearchData.dwr). Scrapable in
  principle, brittle in practice; job rows did not render into readable text.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 18:28:45 +02:00
dennisthiessenandClaude Opus 5 cbb713ba7d chore(aws-fde): NO-GO at Phase 0 gate — 69/100 below the fit bar 
User declined at the gate: 69 is not a strong enough fit to spend a package on.
Consistent with revealed behaviour — every application actually submitted scored
77.5-89; nothing below 75 has ever been sent.

No cohort slot consumed (Adjacent stays 1/2). No bullets written, no resume, no
cover letter. Phase 0 analysis retained in the session file in case a
better-fitting AWS req appears.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 13:02:54 +02:00
dennisthiessenandClaude Opus 5 cc109524f8 feat(aws-fde): Phase 0 for AWS Senior Forward Deployed Engineer, Zurich 
Verbatim JD (amazon.jobs 10504263, posted 2026-08-17) retrieved via Playwright
and stored with the session.

Evidence Fit 69/100, Adjacent. All four Basic Qualifications are Direct, so the
minimum-qualification gate passes. Disclosed rather than buried: forward
deployment is the title-defining capability and is a Gap - SW-4, VZ-1 and
global_forbidden_output_patterns all forbid "customer-embedded delivery", and
application_strategy lists strategic-account FDE as a Stretch target. Scored
Adjacent because AWS's published minimum bar is pure software engineering and
the org is seven weeks old, hiring builders at volume.

Assets: AWS is the evidenced cloud (claims.json forbids GCP output), and Kiro -
named in the JD - is in the verified GenAI toolchain.

CL decision YES. Would consume the last Adjacent cohort slot.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 12:58:39 +02:00
dennisthiessenandClaude Opus 5 e09437dc27 chore(pipeline): close SDU + Google FDE III, triage 5 JDs, drop Louis Dreyfus 
Pipeline:
- SDU PhD: closed, not proceeding. Prof. Jørgensen never replied to the
  2026-08-02 enquiry; user dismissed it two days before the 2026-08-20 deadline.
- Google FDE III GTM: closed, not proceeding after two Google rejections. The
  bar for a third attempt is now a very strong fit; this was Adjacent with an
  agent-orchestration gap. Frees the Adjacent cohort slot (back to 1/2).

JD triage — 5 pulled and read, 1 survives:
- SHORTLIST AWS Senior FDE Zurich: all four basic quals pass and AI/agentic is
  preferred-only, unlike the Google Senior Staff FDE where vector DB + RAG were
  minimum quals. AWS is the evidenced cloud; JD names Kiro, which he uses.
- SKIP AWS ProServe Bern (clinical pharma domain gate), AWS Principal Delivery
  Consultant (8+ yrs architecture leadership + VP+ exec comms), GitLab FDE EMEA
  (Ruby/Go hard gate), MS 200044133 (title says FDE Data Scientist, quals are
  BGP/MPLS/SD-WAN network engineering — mislabelled req).

Roster:
- Dropped Louis Dreyfus: 300 roles, all Brazil/India/Bulgaria ops, zero Swiss.
- Kept MET Group: its "0 eligible" was a filter artifact; exclusion mode reveals
  real Baar/Zug commodity-trading roles.
- Kept Equinor/Telenor/NATO. Memory corrected: the Norway lane is selective, not
  closed — the user still wants very good Norwegian roles.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 12:16:45 +02:00
dennisthiessenandClaude Opus 5 3bb6baa1b3 chore(scout): remove temp probe output files 
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 11:17:40 +02:00
dennisthiessenandClaude Opus 5 aad69745e7 chore(scout): remove temp probe scripts committed by mistake 
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 11:17:34 +02:00
dennisthiessenandClaude Opus 5 284407cd23 fix(scout): repair Roche/Apple, add Amazon+Axpo, make title filtering fail-open 
Scraper fixes:
- Roche: new fetch_phenom adapter (Phenom refineSearch). The old playwright scrape
  of ?locationsearch=Switzerland harvested recommendation-widget cards (Shanghai,
  Kyiv, Bogota) while the page reported no-results. 0 -> 88 CH-eligible roles.
- Apple: dropped default_location "Switzerland", which relabelled US "Various
  Locations" postings as Swiss (84 phantom CH rows over 4 runs). Now honestly 0.
- Meta: NOT broken — metacareers reports "1 Items" for Zurich. Comment added so it
  is not "fixed" again.

New boards:
- Amazon/AWS (fetch_amazon): 32 CH roles incl. a Zurich AWS FDE req and a Bern
  ProServe Cloud Architect. AWS is the evidenced cloud; claims.json forbids GCP.
- Axpo (teamtailor via base_url + pagination): 461 roles, opens the energy lane.
  Locations read from schema.org jobLocation with ISO alpha-2 expanded, so
  Madrid/Milan/Warsaw roles are not marked Swiss. Telenor benefits too.

Title filtering now has two explicit modes. Inclusion allowlists fail closed and
hide unanticipated good-fit roles, so they are now used only where volume forces
it (>~200 roles). Everything else uses the shared, board-agnostic
NOISE_TITLE_EXCLUDE, which fails open and leaves the final call to the scorer and
the reviewer. Palantir stays unfiltered per its existing documented rationale.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 11:17:24 +02:00
dennisthiessenandClaude Opus 5 082a1d250c docs(scout): roster review — broken scrapers, dead lanes, verified adds 
Analysis of the 34-company roster against 4 recent runs and 253 decisions.
No COMPANIES changes.

Findings: Roche/Meta/Apple silently scrape garbage (non-CH rows, 1 intern,
internships-only) — same silent-failure class as the dbt Labs 404; seven boards
have no _title_filter; MET+LDC yielded 0 eligible across 1,464 scrapes.

Verified adds: Amazon/AWS (32 CH roles incl. a Zurich AWS FDE req, and AWS is
the evidenced cloud where GCP is forbidden) and Axpo (Teamtailor custom domain,
400 roles incl. Forward Deployed AI Engineer, opens the named energy lane).

Verified negative: Coinbase proper has 2 EMEA-remote non-eng roles; the
remote-EU data-infra tier is geo-banded below the comp bar.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 10:46:23 +02:00
dennisthiessenandClaude Opus 5 c206563b59 chore(scout): drop Coinbase Ventures getro board 
The board never carried Coinbase's own roles — only Getro portfolio companies.
Across ~3 months it surfaced 3 distinct CH-eligible roles, all from Ashby
(recruiting software, ruled out 2026-07-28) and all decided "skip". Since the
Ashby exclusion it has returned 0 jobs on every run, surfacing a false error
line in each report.

Entry replaced with a "Dropped:" note per the file's convention. fetch_getro is
retained (no caller) since it generalises to any Getro collection id.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 10:34:58 +02:00
dennisthiessenandClaude Opus 5 7b45d41ed8 fix(scout): repoint dead dbt Labs board at Fivetran after merger 
The "dbtlabsinc" Greenhouse board started returning 404, silently dropping
a target company from every scan. getdbt.com/careers now redirects to a page
serving Fivetran's Greenhouse listings (gh_jid links) — the two companies
merged, so the roles live on board "fivetran" (239 jobs, 10 title matches).

Renamed the entry id dbtlabs -> fivetran and cleared the stale seen-state key.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
 2026-08-18 10:30:35 +02:00
48 changed files with 5327 additions and 182 deletions
Expand all files Collapse all files
.gitignore
+4 -1
View File
@@ -24,8 +24,11 @@ __pycache__/
job_scout/.venv/
job_scout/reports/
job_scout/state/*
# ...but track the decision log (job application history), not the churny seen-state
# ...but track the durable application records, not the churny seen-state:
# decisions.json — per-posting decision log (applied/skip/shortlist/closed)
# application_cohort.json — cohort slot tracker (core/adjacent/stretch mix)
!job_scout/state/decisions.json
!job_scout/state/application_cohort.json
# One-off job-board data pulls (debug artifacts)
*_jd.json
AGENTS.md
+3
View File
@@ -157,6 +157,9 @@ _Update this section when starting/finishing a JD._
| Session | Status | Next Command |
|---------|--------|-------------|
| SBB - **Data Engineer (m/w/d)**, Asset Management Infrastrukturanlagen, **Bern** (Job ID 103755) | ⭐ **INTERVIEW INVITED 2026-08-28 - 9. September 2026, 08:30, 45 Min, MS Teams. The only live interview in the log.** Stage **2 of 4** per the posting: *"Virtuelles Kennenlernen mit HR und Fuehrungskraft"*, so **HR together with Andri Wienandts**, not a pure HR screen. Brief: `output/SBB_DataEngineer_AssetMgmt/interview_brief_sbb_2026-09-09.md`. **First conversion of the evidence-first cohort** (6 applications, 2 rejections) and it came from a **COLD submit** - the warm contact printed on the posting was never called. The resume never contained **Snowflake, dbt or Power BI**, the literals an ATS screen would key on, and it still got through: a human read it and the honest-substitution framing held. One data point, do not over-read it. Submitted 2026-08-25, Core, Evidence Fit **79/100**, hard gate PASS, Document Quality **92/100**, no cover letter (SBB waives it). Best practical fit in the log: Bern, German-language, no relocation, EU + B permit. **Four unresolved items, now live call topics:** Anforderungsniveau **K** band (capped GAV, not public, likely below the 180k bar); the seat reads **lateral or below** current Staff + Component Owner scope (the declined-BKW shape); **RAMSI** Java/Spring Boot/Angular is a full responsibility line on a historical/absent stack; Kidz Care scales on gross household income. Gaps to answer honestly: Snowflake, dbt, Argo, Helm, Power BI, Spring Boot, Angular - survivable only because the JD says "z. B.". Thesis stays a **methods prototype**: no operational data, no accuracy figures, PSO surveyed only. | **Decide before 9.9.:** what he does if K lands at 130-150k, if architecture is set elsewhere, or if RAMSI is heavily Angular. Any one can be a no - better decided in advance than improvised on camera. |
| RUAG C5I - **AI Engineer C5I**, **Thun** (application ID 18027) | **SUBMITTED 2026-08-27** via jobs.ruag.ch (portal only). Evidence Fit **74/100**, fit class **Stretch**, hard gate **FAIL** (R1/R2, the title-defining responsibilities, are Adjacent not Direct), Document Quality **93/100** after a critique-and-fix round, channel **Weak/cold**. German CV 2 pages/13 bullets + motivation letter 1 page/295 words; validators PASS, 0 boxes, visual QA clean. **Cohort now 6/10 and the sole Stretch slot is CONSUMED (Stretch 1/1).** Positioning is his canonical title Staff Data, Analytics & AI Engineer; `PP-3` carries current Linux/hardening evidence and `SW-5` ties DevSecOps to C5I's stated model. The letter names the AI-platform gap in one sentence. Honest gaps: no AI/LLM platform, no compute/GPU cluster, no RAG/retrieval, ML/AI frameworks certification-context only. **Compensation, PSP/project eligibility and level were never resolved and are now live screening topics.** Marco Heinzen's published direct line was never used. **If rejected, do not drop RUAG** - Senior DevOps Engineer C5I and Data Lakehouse / Senior Data Platform Engineer are shortlisted and fit better. | Await response; optional calls to Marco Heinzen (scope/level) and Frank Haugwitz (comp/PSP). |
| Schweizer Armee - **Kommando Cyber (Kdo Cy)**, DevOps Engineer III (Data Platform), **Zimmerwald** (Ref JRQ$540-19848) | **SUBMITTED 2026-08-27.** Evidence Fit **79/Core**, hard gate **PASS**, Document Quality **92/100**, channel **Weak**. German 2-page CV plus 1-page motivation letter; validators PASS, 0 boxes, visual QA clean. BW-1 surfaces the six-year Bundeswehr officer career. User explicitly accepted the critique's remaining BS-1 wording issue (*„ohne manuellen Eingriff“*) as good enough and submitted unchanged. Cohort evidence-first-2026-01 is now 5/10, Core 3/7. | Await response; optional post-submit call to Marcel Matthey-Doret about Lohnklasse, Engineer-III level and development path. |
| Google - Software Engineer III, Business Home, Zurich (req 93922217108087494) | **CLOSED - NOT PROCEEDING 2026-07-28** (submitted 2026-07-27; Core, Evidence Fit 89/100, Document Quality 94/100; no interview). Rapid early-screen disposition; known risks were Mid-level versus current Staff scope, data/platform versus product-SWE positioning, preferred algorithms/accessibility gaps, and a cold channel. | Done - cohort rejection recorded; do not treat this outcome alone as a document-quality failure |
| Microsoft — Principal Forward Deployed Engineer, SWE (German Speaking), Zürich (req 200043897) | **SUBMITTED 2026-07-27** (84.2/100 Pass 2; finalized 2-page resume + 1-page cover letter). Strong native-German, Staff progression, production ownership and enterprise-data-readiness case; honest gaps remain in end-to-end LLM delivery, Azure AI and strategic-account FDE experience. | Done — await response |
| BIS Basel — Senior Data & Analytics Engineer, AI (jr100429, 3-yr term) | **SENT 2026-07-10** (80.5/100; deadline 2026-07-24; hybrid Basel, English-working intl org). Critique flags LLM-depth probe (integration/config vs. ownership) as the screening risk — prep honest answer before any call | Prep interview brief when screening lands |
CLAUDE.md
+23 -5
View File
@@ -157,9 +157,14 @@ _Update this section when starting/finishing a JD._
| Session | Status | Next Command |
|---------|--------|-------------|
| Google — Forward Deployed Engineer III, Google Cloud GTM (French, German), Zürich (req 78350205438567110) | **PHASE 0 DONE 2026-08-11** — Adjacent, **Evidence Fit 67/100**, hard gate PASS (all 5 minimum quals hold; German alone satisfies the "French **or** German" requirement), Channel Weak. Level badge is **Mid**; user explicitly overrode level/comp 2026-08-11 (*"ignore fit, i would gladly accept a mid FAANG offer"*) — do NOT re-raise level. **Caveat:** responsibility #1 (lead developer, production agentic workflows / multi-agent / MCP) is a **Gap** — SW-8 forbids agent-orchestration claims; ADK/LangGraph/CrewAI are `never-used`. GCP also unevidenced (AWS-primary). Consumes the **last Adjacent cohort slot (2/2)**. Sibling req *FDE III, Generative AI* confirmed CLOSED. CL decision YES — its job is pre-empting the overqualification screen that killed Business Home in ~24h | Phase 1 — bullet plan (awaiting user confirmation of role type, format, framing) |
| SDU Center for Energy Informatics, Odense DK — PhD, Theme 2 Predictive Maintenance & Asset Management of Smart Energy Networks (job 4159) | **EXPLORATORY 2026-08-02** (deadline 2026-08-20). Real JD scraped via Playwright. Strong genuine fit: thesis = vibration condition monitoring, RBR+ANN hybrid, throughput-vs-sensor-rate finding; Bosch fab sensor data; Swisscom data products as AI foundation. **Hard constraint: DKK 37,075/mo ≈ CHF 56k — far below the 180k bar; user accepts this knowingly as a PhD.** Real blocker is 2 letters of recommendation (13 yrs out of academia). Top-30% doc does NOT apply (numeric grading) — do NOT submit the ECTS-B/top-35% certificate against it. No publications. **Enquiry email to Prof. Bo Nørregaard Jørgensen SENT 2026-08-02** (draft in `output/sdu_phd_energy_informatics/enquiry_email_joergensen.md`); asked two questions — industry-candidate viability, and whether non-academic letters of recommendation are acceptable. | Await reply. Critical path is the 2 LORs, not the documents — chase referees regardless of reply. Build an **academic-style CV** (education-first, not the 2pp industry resume) when the user gives the go-ahead |
| Aker BP ASA — Data Product Architect, AI-ready Data Products (FINN 469067315) | **SUBMITTED 2026-07-30** (ahead of the 2026-08-02 deadline; Adjacent, **Evidence Fit 79/100**, Document Quality 93/100, hard gate PASS, Channel Weak; Stavanger/Oslo/Trondheim, English working language, no Norwegian and no clearance required, EEA work rights). 2pp resume + 1pp CL, both validator PASS. Honest practitioner framing: builds governed data products *inside* Swisscom's Data Mesh; role *authors* an enterprise framework — real level stretch, never fabricate authorship. **Atlassian Compass closed the catalogue-tool gap** (74→79). Remaining gaps: framework authorship, no energy domain, AWS vs their Microsoft/Cognite stack, **zero verified metrics**. | Done — await response; optional follow-up email to Per Olav Marthinsen |
| RUAG C5I - **AI Engineer C5I**, **Thun** (application ID 18027) | **SUBMITTED 2026-08-27** via jobs.ruag.ch (portal only; email and post refused). Evidence Fit **74/100**, fit class **Stretch**, hard gate **FAIL** (R1 and R2, the two title-defining responsibilities, are Adjacent not Direct), Document Quality **93/100** after a full critique-and-fix round (85 pre-edit), truth and provenance 24/25, **0 Tier 1 findings at Round 2**, channel **Weak/cold**. Submitted: German CV **2 pages, 13 bullets** + German motivation letter **1 page, 295 words**; both validators PASS, 0 boxes, text-order and visual QA clean, submission PDFs MD5-verified before sending. **Cohort evidence-first-2026-01 is now 6/10 and the sole Stretch slot is CONSUMED (Core 3/7, Adjacent 2/2, Stretch 1/1)** - any further application must clear Core or Adjacent on its own merits. Positioning is his canonical title **Staff Data, Analytics & AI Engineer**, carried by `BS-1` production ML-inference integration, `BS-6` Linux/Ansible, `SW-3`/`SW-7`/`SW-2` current Kubernetes and data-product work, `PP-3` current self-hosted Linux and hardening, `SW-5` DevSecOps tied to C5I's stated operating model, and the canonical `BW-1` officer career. The letter **names the AI-platform gap in one sentence** and pivots to the operations/platform side. **Honest gaps that travel with it:** no AI/LLM platform built or operated, no compute/GPU cluster, no RAG/retrieval implementation, no formal LLM evaluation; ML/AI-framework depth is certification-context only. **Compensation, PSP/project eligibility and role level were NEVER resolved before sending, so all three are now live screening topics rather than pre-cleared ones.** ⚠ **Marco Heinzen's published direct line (+41 79 568 14 96) was never used** - this went in as a pure cold submit, the sixth consecutive one. **If rejected, that is NOT a signal to drop RUAG:** five other RUAG C5I reqs sit in the scout log, and **Senior DevOps Engineer C5I** and **Data Lakehouse / Senior Data Platform Engineer** are both already shortlisted and are **materially better fits than this one**, sitting on the lane where his evidence is Direct rather than Adjacent. | Await response. Optional but still worthwhile: call **Marco Heinzen** about technical scope and level, and **Frank Haugwitz** about compensation and PSP/project eligibility. Do not react to a single outcome with positioning changes (`application_strategy.md` §38/§74). |
| Schweizer Armee - **Kommando Cyber (Kdo Cy)**, DevOps Engineer III (Data Platform), **Zimmerwald** (Ref JRQ$540-19848) | **SUBMITTED 2026-08-27.** Evidence Fit **79/Core**, hard gate **PASS**, Document Quality **92/100**, channel **Weak**. Submitted package: German 2-page CV with **15 bullets** and 1-page motivation letter with **300 body words**; validators PASS, 0 boxes, text order and visual QA clean. Canonical **BW-1** records the six-year Bundeswehr officer career and anchors the letter's military/public-service motivation. User explicitly accepted the critique's remaining BS-1 wording issue (*„ohne manuellen Eingriff“*) as good enough and submitted unchanged. Honest gaps remain physical datacenter work, GPGPU, ClickHouse and named Pull-GitOps practice. Cohort evidence-first-2026-01 is now **5/10, Core 3/7**. | Await response; optional post-submit call to Marcel Matthey-Doret about Lohnklasse, Engineer-III level and development path. |
| Citadel Securities — Platform Engineer, Research Platform Engineering, **Zürich** (also NY/Miami) | **SUBMITTED 2026-08-26.** Cohort slot: **Adjacent 2/2** (cohort now 4/10). ⚠️ **KEY FINDING — the application form's preferred-work-location selector offered NO Zurich**, despite the JD saying "Miami, Zurich or New York" and the site's own Zurich filter returning this req. Careers site is operated by **Citadel Enterprise Americas LLC**; the apply flow is likely US-entity, or the Zurich seat is closed while the posting stands. **This partly supersedes the working-model risk** — if no Zurich seat is reachable through that form, hybrid-vs-onsite is moot and the application may sit against a US req, a hard NO on relocation. **Rule for any future Citadel application: check the form's location selector BEFORE building a package** — the posting's stated locations and the site filter both proved unreliable. User applied with low expectation; cold channel, evergreen standing req. 2pp resume (13 bullets) + 1pp CL (270 words), both validator **PASS, 0 warnings**; JD term coverage 51%→68% after Tier 1 fixes. Critique **STALE by design** (86/100 pre-edit; user chose to submit as-is over re-critiquing; est. ~91 post-edit, never quote as a score). **Cohort entry and decision-log entry deliberately NOT written — no confirmed send date.** Commands ready in the session file. **The working-model question is still open and travels with the application:** Evidence Fit **71/100 (Adjacent, mid-band; revised down from 74)**, qualifications gate PASS but **a practical constraint is UNRESOLVED**, which `critique_framework.md` treats as a NO-GO trigger. **The JD is silent on hybrid vs onsite (0 grep matches), and Ken Griffin is on record that Citadel returned to the office 5 days a week and calls it his most important leadership decision.** If that holds for Zurich, it means ~2–2.5h daily commuting from Bern or relocating — breaking the Bern-based / 2–3 day hybrid bar. Only contrary evidence is a stale 2022 anonymous Fishbowl post about Operations. **Must be asked, not researched** — same class as the SBB K band. Practical constraints scored 5/5 in Phase 0 on an unexamined assumption; corrected to 2/5. **Zero Gaps among the 7 minimum quals**, and all five preferred technologies (Kafka, Kubernetes, Spark, Airflow, distributed DBs) are Direct — **the strongest raw stack alignment in the log**. Title-defining function **splits**: R2 (ingestion/transformation/storage/lifecycle of large datasets) is Direct and describes SW-1/SW-7 exactly; **R1 (design and build the research platform itself) is Adjacent** — the Aker BP authoring-vs-building shape again, and unwritable-around under Scope Discipline. **Classification caveat:** `application_strategy.md`'s "a defining responsibility is only Adjacent → Stretch" would strictly class this Stretch; recorded Adjacent per the Aker BP precedent. Both cohort slots free either way. Gaps: platform authoring at firm scale, no professional quant/finance (now partly mitigated by new **PP-1/PP-2**), SDKs (0 evidence), low-latency/HFT (0 evidence — thesis throughput figure is explicitly barred by `thesis_limits`). **Go is NOT a gap** — Q3 is disjunctive and Python satisfies it. **$175–350k in the posting is a US/NY pay-transparency disclosure, NOT the Zurich number — must be asked.** Channel **Weak/cold** — would be the 5th consecutive cold application, in an office visibly staffed from Google and ETH (platform research head Costas Bekas; Nicolai Meinshausen leads principal research). Cover letter: **YES** if it proceeds — the PP-1 domain story is the main differentiator and needs prose. | **Ask Citadel Securities recruiting whether the Zurich seat is hybrid or 5-day onsite BEFORE building the package.** If 5-day onsite → likely NO-GO on the Bern constraint. If hybrid 2–3 days → resume Phase 2 (plan is written and budget PASSes at 11–14 bullets; two open items: Vizrt VZ-1 C++/distributed bullet, and adding a Projects section for PP-1) |
| SBB - **Data Engineer (m/w/d)**, Asset Management Infrastrukturanlagen, **Bern** (Job ID 103755) | **INTERVIEW INVITED 2026-08-28 - 9. September 2026, 08:30, 45 Min, MS Teams.** Per the posting's own process this is **stage 2 of 4**, *"Virtuelles Kennenlernen mit HR und Fuehrungskraft"* - **HR together with Andri Wienandts**, not a pure HR screen. Brief written to `output/SBB_DataEngineer_AssetMgmt/interview_brief_sbb_2026-09-09.md`. **This is the FIRST interview of the evidence-first cohort** (6 applications, 2 rejections, 1 conversion) and it converted from a **COLD submit** - the channel plan called for phoning Wienandts first and that never happened. **The resume never contained Snowflake, dbt or Power BI**, the three literals an ATS keyword screen would have keyed on; the session had recorded that risk as "accepted, not solvable". It did not filter him out, so a human read the dossier and the honest-substitution framing survived first contact. **One data point, do not over-read it** - but it is evidence against assuming honest omissions are fatal at the screen. Submitted 2026-08-25: 2pp English resume, 13 bullets, no cover letter (SBB waives it), Evidence Fit **79/100 Core**, hard gate **PASS**, Document Quality **92/100**, channel **Weak/cold**. Best practical fit in the log - Bern-based, German-language, no relocation, EU citizen + B permit. **Four things unresolved before submitting and now live call topics:** (1) **Anforderungsniveau K** is a capped GAV band, figures not public, very likely below the 180k bar; (2) the seat reads **lateral or below** current Staff + Component Owner scope, the same shape as the declined BKW, and the JD's only architecture language is a personal skill not an ownership mandate; (3) **RAMSI** Java/Spring Boot/Angular is a whole responsibility line on a stack that is historical or absent - ask its share, do not oversell; (4) Kidz Care scales on gross household income, so far below 90% at this band. Named gaps to answer honestly: Snowflake, dbt, Argo Workflows, Helm, Power BI, Spring Boot, Angular - all non-canonical, all survivable only because the JD prefixes the stack with "z. B.". Thesis stays a **methods prototype**: no real operational data, no accuracy figures, PSO surveyed only, throughput figure barred. | **Prep before 9.9.:** decide in advance what he does if K lands at 130-150k, if architecture is set elsewhere, or if RAMSI is heavily Angular - any one can be a no, and it is better decided before the call than improvised on camera. |
| AWS (AWS EMEA SARL, Switzerland Branch) — Senior Forward Deployed Engineer, Zürich (job 10504263) | **CLOSED — NO-GO 2026-08-18** at the Phase 0 gate. Evidence Fit **69/100** (Adjacent). Minimum-qual gate passed (all 4 Basic Quals Direct) but the *title-defining* capability — customer embedding — is a Gap that `claims.json` globally forbids claiming, and R2 multi-agent/retrieval is a second Gap. User declined: **69 is below the fit bar worth spending a package on.** Consistent with revealed behaviour — every application actually submitted scored **77.5–89**; nothing below 75 has ever been sent. **No cohort slot consumed (Adjacent stays 1/2).** Phase 0 analysis retained in the session file for reuse if a better-fitting AWS req appears. | Done — no package built |
| Google — Forward Deployed Engineer III, Google Cloud GTM (French, German), Zürich (req 78350205438567110) | **CLOSED — NOT PROCEEDING 2026-08-18.** Phase 0 was done (Evidence Fit 67/100, hard gate PASS) but user declined to proceed after **two Google rejections** (Merchant Data Science, Business Home): the bar is now that a Google req must fit *very* well to be worth a third attempt, and this one was Adjacent with an agent-orchestration Gap. **Frees the Adjacent cohort slot (back to 1/2).** Phase 1 never started. | Done — do not reopen without a materially stronger Google req |
| SDU Center for Energy Informatics, Odense DK — PhD, Theme 2 Predictive Maintenance & Asset Management of Smart Energy Networks (job 4159) | **CLOSED — NOT PROCEEDING 2026-08-18.** Enquiry email to Prof. Bo Nørregaard Jørgensen (sent 2026-08-02) went unanswered for 16 days; user dismissed the idea. Genuine thesis fit (vibration condition monitoring, RBR+ANN hybrid) but the blockers were never the documents — 2 letters of recommendation 13 years out of academia, and DKK 37,075/mo ≈ CHF 56k against a 180k bar. Deadline was 2026-08-20. | Done — no further action |
| Aker BP ASA — Data Product Architect, AI-ready Data Products (FINN 469067315) | **CLOSED — REJECTED 2026-08-27** (submitted 2026-07-30, ~28 days to disposition, no interview; the session file and cohort record record the submission as 2026-07-29 — a one-day inconsistency, never reconciled). Adjacent, **Evidence Fit 79/100**, Document Quality 93/100, hard gate PASS, **Channel Weak/cold** — the optional follow-up email to Per Olav Marthinsen was left unsent, so this went in as a pure cold submit. Stavanger/Oslo/Trondheim, English working language, no Norwegian and no clearance required, EEA work rights. 2pp resume + 1pp CL, both validator PASS, no Tier 1 truth findings — **a cold-channel no-interview rejection does not implicate document quality on its own** (same disposition class as Microsoft ISE and Google Business Home). Honest practitioner framing: builds governed data products *inside* Swisscom's Data Mesh; role *authored* an enterprise framework — real level stretch, never fabricated authorship. **Atlassian Compass closed the catalogue-tool gap** (74→79). Standing gaps if a similar req appears: framework authorship, no energy domain, AWS vs their Microsoft/Cognite stack, **zero verified metrics**. **Cohort slot stays consumed** (evidence-first-2026-01, Adjacent 1/2 — a rejection does not free a slot; cohort remains 4/10). **The Norway lane stays OPEN and selective** — NATO JWC Stavanger is still live, and Equinor/Telenor and other strong Norwegian employers remain in scope per the user's standing instruction. Never frame this lane as closed. | Done — cohort rejection recorded; do not react to this single outcome with positioning changes (`application_strategy.md` §38/§74) |
| Google - Software Engineer III, Business Home, Zurich (req 93922217108087494) | **CLOSED - NOT PROCEEDING 2026-07-28** (submitted 2026-07-27; Core, Evidence Fit 89/100, Document Quality 94/100; no interview). Rapid early-screen disposition; known risks were Mid-level versus current Staff scope, data/platform versus product-SWE positioning, preferred algorithms/accessibility gaps, and a cold channel. | Done - cohort rejection recorded; do not treat this outcome alone as a document-quality failure |
| Microsoft — Principal Forward Deployed Engineer, SWE (German Speaking), Zürich (req 200043897) | **SUBMITTED 2026-07-27** (84.2/100 Pass 2; finalized 2-page resume + 1-page cover letter). Strong native-German, Staff progression, production ownership and enterprise-data-readiness case; honest gaps remain in end-to-end LLM delivery, Azure AI and strategic-account FDE experience. | Done — await response |
| BIS Basel — Senior Data & Analytics Engineer, AI (jr100429, 3-yr term) | **SENT 2026-07-10** (80.5/100; deadline 2026-07-24; hybrid Basel, English-working intl org). Critique flags LLM-depth probe (integration/config vs. ownership) as the screening risk — prep honest answer before any call | Prep interview brief when screening lands |
@@ -167,7 +172,7 @@ _Update this section when starting/finishing a JD._
| Microsoft — Senior SWE, Industry Solutions Engineering (ISE), Zürich (req 200040836) | **CLOSED — REJECTED 2026-08-11** (submitted 2026-07-03, 85.8/100 Pass 2; 2pp resume + 1pp CL; verbatim Eightfold JD; IC4 base CHF 146.2–245.9k; applied ~7 days after posting). No interview, ~39 days to disposition. Cold-channel application; does not implicate document quality on its own. The separate Microsoft **Principal FDE** req 200043897 is a different pipeline and remains open | Done — cohort rejection recorded |
| Kraken (Payward) — SRE, AI Agents (remote, CH-eligible) | **CLOSED — REJECTED 2026-06-17** (applied 2026-06-15 ~87.2/100, no interview). Honest gaps (NO Terraform/SRE-title/LangGraph) likely the filter; 4th Kraken req declined/rejected to date | Done |
| Google — Senior Data Engineer (Merchant Data Science), Zürich/MV | **CLOSED — NOT PROCEEDING 2026-07-24** (applied 2026-06-15, 85.5/100; passed Google Hiring Assessment 2026-06-20, no interview). Assessment pass remains recorded separately. The official 90-day wait applies only to reapplying for the same job; different Google roles remain eligible, subject to 3 applications per rolling 30 days. | Done — target a different strong-fit Google req |
| Snowflake — Sr SWE, Enterprise (Observe by Snowflake), Zürich | **SENT 2026-06-06** (~86/100; 2pp resume + 1pp CL; real Ashby JD; comp CHF 176–253k base; NO C++ gate). Tier 1+2 applied; Vizrt low-latency skipped per user. Best-fit role in the 2026-06 search | Done — await response |
| Snowflake — Sr SWE, Enterprise (Observe by Snowflake), Zürich | **CLOSED — NO RESPONSE 2026-08-25** (sent 2026-06-06, ~86/100; 2pp resume + 1pp CL; real Ashby JD; comp CHF 176–253k base; NO C++ gate). Closed by user decision at **80 days silent** — past the longest real disposition in the log (Equinor, 79 days). **No rejection was ever received: this is a presumed-dead call, not a rejection — do not count it in rejection statistics or treat it as a document-quality signal.** Cold channel. Pre-dates the evidence-first-2026-01 cohort (started 2026-07-27), so **no cohort slot freed**. Highest-scoring package in the log at ~86 and it drew no reply at all — the strongest single argument in the log for the channel problem over the document problem. **Unblocks the Snowflake Observe Metrics Platform req** (`db4f0492`), paused 2026-07-28 solely to avoid stacking a second cold application on the same org while this one was open | Done — if Snowflake is still wanted, the Metrics Platform req is now open to pursue, but route it through a recruiter rather than another cold submit |
| Isovalent (Cisco) Sr Data Engineer, Observability | **CLOSED — role pulled** (live Cisco scrape 2026-06-02: not on board; Recruitee link dead). Package finalized ~86/100, SHELVED for reuse | Done — retarget PDFs to next live data-eng req (QuantCo/Grafana/Confluent) |
| Google Zürich Sr SWE Infrastructure (Data Pipeline) | **CLOSED — DROPPED + DELETED 2026-06-02** (poor fit). Live JD = Core infra/systems SWE with **C++ as a MINIMUM qual**, off-thesis vs `user_positioning`. Output folder deleted (was built on a fabricated JD). | Done — do not reattempt this req |
| Kraken AI Infrastructure | **CLOSED — REJECTED** (applied, no interview) | Done |
@@ -175,7 +180,7 @@ _Update this section when starting/finishing a JD._
| Infineon AI Engineer | **CLOSED — not applied** (no relocation to Germany) | Done |
| Apple Data Engineer (ISE, Zurich) | **CLOSED — REJECTED** (no interview) | Done |
| Google FDE GenAI (Zurich) | PAUSED — GenAI evidence gap too large; redirecting to data-eng/MLOps roles | Likely abandon |
| Equinor — Manager / AI Architect, Agentic systems (Stavanger/Oslo/Rotvoll/Sandsli; req JR106747) | **CLOSED — REJECTED 2026-08-14** (submitted 2026-05-27, ~80/100; posting deadline 05.06.2026). No interview, **79 days** to disposition — the longest turnaround recorded. Cold channel. This was a **Manager** req, not a pure IC architect role, so level/scope stretch compounds the standing Norway constraints (comp below bar, Norwegian A2, no clearance). **Pre-dates the evidence-first-2026-01 cohort (started 2026-07-27) — does not free a cohort slot.** Norway lane now reduced to the open Aker BP application. | Done — cohort-external rejection recorded |
| Equinor — Manager / AI Architect, Agentic systems (Stavanger/Oslo/Rotvoll/Sandsli; req JR106747) | **CLOSED — REJECTED 2026-08-14** (submitted 2026-05-27, ~80/100; posting deadline 05.06.2026). No interview, **79 days** to disposition — the longest turnaround recorded. Cold channel. This was a **Manager** req, not a pure IC architect role, so level/scope stretch compounds the standing Norway per-role gates (norsk working-language, no NO clearance — **not** comp: the 180k bar is CH-only and does not apply to Norway). **Pre-dates the evidence-first-2026-01 cohort (started 2026-07-27) — does not free a cohort slot.** **The Norway lane stays OPEN but selective** — Equinor, Telenor and NATO JWC remain in the scout; the user has explicitly said he still wants to reach out for very good Norwegian roles, Equinor included, after this rejection. Never frame this lane as closed. | Done — cohort-external rejection recorded |
| QuantCo Cloud Engineer (Europe/Zürich) | **CLOSED — REJECTED** (applied 2026-06-01 ~82/100, no interview; rejection 2026-06-15) | Done |
---
@@ -186,9 +191,22 @@ _See `config.md` for user-specific corrections. Add verified errors here as you
| Date | Correction | Files fixed |
|------|-----------|-------------|
| 2026-08-27 | **A German-language resume was being typeset with English hyphenation patterns, and had been for every German package built from this template.** The RUAG resume loaded `lmodern` but **never `babel`** - the cover letter did (`\usepackage[ngerman]{babel}`), the resume did not. The compiled PDF broke words as `Hal-bleiterfertigung`, `Tran-skription`, `automa-tisierte` and `Foun-dation`: all wrong in German, and exactly the kind of thing a native-German Swiss reader notices without being able to name. It survived an earlier full visual QA because nothing *looks* broken - the lines are justified and the page is clean. Adding `\usepackage[ngerman]{babel}` produced correct breaks (`Halb-leiterfertigung`, `automati-sierte`) and **also cleared an overfull box** introduced when the headline grew. ngerman still breaks `Transkription` after `Tran`, so `\hyphenation{Trans-krip-ti-on}` was added. **Rule: any German-language document must load `babel[ngerman]`, and hyphenation must be checked in the compiled PDF, not assumed from a clean box count.** `resume_template.tex` should be checked for the same omission. | RUAG resume (fixed); `resume_template.tex` (OPEN) |
| 2026-08-27 | **`IBM AI Engineering` was on resumes for months with no primary-source record, and `PyTorch` was under-recorded as a result.** The `/critique` flagged it as the only line a canonical preflight could not confirm: the other three certifications sit in `thiessen_certifications.md` with issuer, date and certificate number, while this one existed only in `bundle_ml_ai_engineer.md`. The user supplied the PDF (`C:\myCloud\Bewerbungsunterlagen\Zeugnisse\Zertifikate\cert_IBM_AI_Engineering.pdf`) and it was read directly: **IBM via Coursera, 4 Jun 2020**, no expiry, verify code `3ZBZFVAL6A34`, six courses, name on the certificate `Dennis Thiessen` (sz variant, as on ITIL and iSAQB). Recorded as entry **#8**. It also proves **PyTorch**, which `claims.json` had as `coursework-or-personal-unverified`; upgraded to `evidence: certification` with `output` unchanged at `certification-context-only`. The certificate is explicitly **non-credit** and confers no grade or degree - never present it as an academic qualification. **Same shape as the Ansible/Linux/Bosch-promotion gaps: the claim was true, the KB simply had no record of it.** | `thiessen_certifications.md` (entry 8); `claims.json` (PyTorch evidence) |
| 2026-08-27 | **`claims.json` had no Bosch promotion date, so a Swisscom-style title line was unwritable from the canonical file.** Asked to render Bosch like Swisscom ("Senior Data Engineer, Data Analysis (Beförderung xy)"), the canonical file offered only a combined `official_title` "Engineer / Senior Engineer, Data Analysis" with **no `title_history` and no date** - while SWISSCOM has a full `title_history`. The data existed one layer down: `experience_bosch.md` records it LinkedIn-confirmed as Data Engineer **Feb 2020 - Jan 2021** and **Senior** Data Engineer **Jan 2021 - Dez 2022**. Added to `claims.json` as `title_history`, with `display_title` set to the user's own preferred wording **"Senior Data Engineer, Data Analysis"** and `official_title` left untouched as the formal Zeugnis string. **Systemic lesson, the same shape as the Ansible/Linux gaps: `claims.json` is the declared highest authority but is not always the most complete file - when it is silent, check the `experience/` files before telling the user something is unwritable.** Edit note: the file uses compact inline JSON and CRLF; a `json.dumps(indent=2)` rewrite reformats all ~700 lines. Patch it as text. | `claims.json` (BOSCH title_history, display_title); RUAG resume |
| 2026-08-27 | **Vizrt shown as "DevOps Engineer" alone, on user instruction.** Canonical `official_title` is **"Test Automation Engineer"**; `display_title` is "Test Automation / DevOps Engineer". The user asked to drop "Test Automation" for the RUAG C5I package. Bullet content (CI/CD quality gates, Python backend for distributed transcoding) supports the DevOps framing and the user is the authority on his own role - but **this is the one title on that document that will not match a Zeugnis word-for-word**, so flag it rather than silently propagate it. Not written into `claims.json`: it is a per-package display choice, not a canonical correction. | RUAG resume (session file records the flag) |
| 2026-08-27 | **A finished package had no reviewable PDF, and the only PDF that existed was a stale pre-fix build.** The RUAG C5I resume was recorded as complete, but the previous session had compiled only into a scratch folder (`.tmp_ruag_ai_resume/`), so `output/RUAG_AI_Engineer_C5I/` held just the `.tex`. The user could not approve what he could not open. Worse, that scratch folder still contained a **pre-correction PDF that included Capgemini**, named with the deliverable's stem - one grab away from being submitted. The folder also held a stale 3-page text extraction, which is why the prior session's "both pages visually inspected" claim had been made against a different document than the final one. **Rule: compile the PDF into the application's own `output/` folder** (as the Kdo Cy package does), and never leave a superseded PDF beside a deliverable. "Output .tex only" means the .tex is the source of truth, **not** that the user should be handed nothing to look at. | RUAG output folder (PDF added, stale temp folder deleted); `config.md` (Output Rules) |
| 2026-08-27 | **Capgemini reappeared in a generated resume, breaking a standing omit preference.** The RUAG C5I draft carried `Capgemini Deutschland GmbH, Nov. 2014 - Mai 2015` as a compact chronology entry. The user does **not** want Capgemini on any document (six months only; short-stay impression in Germany). A corpus check found it was the **only** generated `.tex` under `output/` containing the name, so this was a one-off regression, not a systemic drift. **Root cause: the preference had no first-class home.** It lived in memory and in a parenthetical inside the unrelated "12+ years" entry below, so a fresh model reading `CLAUDE.md` and `config.md` alone would not see it - `config.md` does not mention Capgemini at all. Recorded here as its own rule: **Capgemini is omitted from every resume, CV and cover letter. The resulting Nov. 2014 - Mai 2015 gap is intentional and must never be "fixed".** The visible chronology starts at Generali, Mai 2015 - which is also what the "over ten years" profile wording is measured against. | RUAG resume (fixed); this log (rule recorded) |
| 2026-08-27 | **The English resume template asserts "12+ years of production software experience" and that does not hold against `claims.json`.** Canonical `employment` starts at **Capgemini 2014-11** → ~11 years 9 months as of Aug 2026; and Capgemini is omitted from documents by standing user preference, so the *visible* record on any generated resume starts **Generali 2015-05** → ~11 years 3 months. Either reading falls short of twelve. Inherited unexamined into the Kdo Cy draft's opening sentence and caught at final review; corrected there to **"über zehn Jahren"**, which is true under both readings. **`resume_builder/templates/resume_template.tex` still carries the 12+ claim and will re-infect the next package.** Possible innocent explanation: officer service after the 2013 M.Eng. (he is a Universität-der-Bundeswehr graduate) may be counted and is not in `employment` — **ask before editing the template**, and if it counts, record it in `claims.json` so the number has a source. Lesson: a number inherited from a template is still a claim and needs a canonical source like any other. | Kdo Cy resume (fixed); `resume_template.tex` (OPEN) |
| 2026-08-27 | **Linux administration was missing from `claims.json` entirely — the second infrastructure blind spot found the same day, and the more consequential one.** User-confirmed: at **Bosch (2020–2022) Linux servers carried everything** — ML platform hosts, Docker hosts, backend services, the Ansible control path — and he administered and automated them; he also **wrote Ansible extensions**, including a credential plugin pulling secrets from a password keystore with local caching to cut lookup volume (scalability/performance). Recorded as new claim **BS-6**. Separately, he has run a **self-hosted Debian server for ~10 years** — nginx, mail, Nextcloud, VPN, Docker services, Bitwarden — which he administers *and hardens*; recorded as **PP-3** with a PP-1-style forbidden list (never enterprise/production/multi-user, no uptime or scale figures, never employer work, never evidence of physical-datacenter experience). **Hardening evidence leans on the personal side — always say so.** **GitOps question resolved and the answer is NO:** playbooks were in Git but execution was **push-based from Jenkins/GitLab**, no reconciliation agent. The word stays `forbidden`; the substance is writable as *"versionskontrollierte Infrastrukturautomatisierung mit Ansible aus Git, ausgeführt über Jenkins-/GitLab-Pipelines"*. **Systemic lesson: the KB is data-engineering shaped and under-records infrastructure and operations work.** Two canonical gaps (Ansible, Linux) cost the Kdo Cy assessment ~15 points and produced an initial NO-GO that reflected the KB, not the candidate. Before scoring any DevOps/platform/SRE-flavoured JD, ask about ops evidence explicitly rather than trusting `claims.json` to be complete. | `claims.json` (skills Linux administration + GitOps note, new claims BS-6 and PP-3) |
| 2026-08-27 | **Ansible was missing from `claims.json` entirely — a real, currently-relevant skill that could not legitimately be listed.** It already appeared in `experience_bosch.md` BS-1 bullet variants ("Docker, Kubernetes, Ansible"), but the canonical file had **zero** mentions in either `skills` or any claim, so by the evidence-first rule it was unwritable. User confirmed 2026-08-27 that Ansible work at Bosch (2020–2022) was **intensive** and went beyond the BS-1 ML-inference orchestration into configuration management and infrastructure automation. Added as `production-historical`, `output: allowed`, and BS-1's scope now records it. **This is genuine IaC evidence** alongside CloudFormation and it materially improves any DevOps/platform JD — it does **not** license Terraform, which stays unverified. **`GitOps` added as a separate, explicitly `forbidden` skill pending clarification:** the user described the Ansible work as GitOps, but Ansible is configuration management while GitOps is narrower (declarative desired state in Git, reconciled by an agent such as Argo CD or Flux). Until it is established whether playbooks lived in Git and were applied from there, and whether a reconciliation agent was involved, the word must not appear in any document. The Kdo Cy JD names **"Pull-GitOps"** explicitly, so that is the one audience guaranteed to probe the distinction. | `claims.json` (skills Ansible + GitOps, BS-1 scope) |
| 2026-08-21 | **The data-engineer bundle carried the SW-1 scope violation it was supposed to prevent.** `bundle_data_engineer.md` §S5's cover-letter opening hook read "*while simultaneously leading the migration of our legacy stack to a cloud-native AWS architecture*" — a full-ownership verb on a company-scale object, directly contradicting `claims.json` SW-1 `forbidden` **and** point 2 of the narrative thread three lines below it. Live trap for every future data-eng cover letter. Rewritten to the scoped form plus an inline scope warning. **Check the other four bundles for the same pattern.** | `bundle_data_engineer.md` (§S5 hook + new warning) |
| 2026-07-27 | **SW-1 was not solo.** Swisscom AWS migration was written as "sole technical lead" / "Led migration of legacy stack." Dennis was primary engineer for **his own domains'** pipelines and a contributor to the wider programme. | `experience_swisscom.md` (role line + all 3 bullet variants + overclaiming warning), `config.md` |
| 2026-07-27 | **Security Champion is 2025/2026 only, and is a team role — not an award.** Source files claimed "3 consecutive years (2023/24–2025/26)." User has now corrected this twice. **Default is OMIT** unless the JD explicitly requires security/DevSecOps. | `experience_swisscom.md` SW-5, `achievement_reframing_guide.md`, `skills_taxonomy.md` (3 rows) |
| 2026-07-29 | **Atlassian Compass is the metadata/lineage platform for Swisscom data products** (user-supplied). Previously the KB named no catalogue/lineage product at all, which read as a hard gap against data-governance JDs. Practitioner use only — never claim admin, rollout or ownership. Does NOT license claiming Purview/Collibra/Alation. | `claims.json` (SW-7 scope + skills entry), `experience_swisscom.md` SW-7 |
| 2026-08-02 | **Master's thesis verified against the PDF.** Vibration-based condition monitoring of CNC machine tools; hybrid **rule-based reasoning + 7-10-3 ANN**; throughput/latency evaluation (~500 SPS pipeline vs 72.9 kSPS sensors). **PSO was surveyed but NOT implemented** — an earlier note in this session wrongly listed it as an applied method. No real operational data and no accuracy figures: it is a methods prototype, not a validated system. Also recorded: ECTS relative grade **B (top 35%)**, English-language transcripts exist. | `claims.json` EDU-MENG |
| 2026-08-26 | **Degree final grades were missing from the KB entirely.** A Citadel application form asked for GPA and neither degree's *Abschlussnote* was recorded anywhere in `claims.json` — only the M.Eng. **thesis** grade (1.0). The only lead was `knowledge_base/extractions/thiessen_cv_master_profile.md`, whose own provenance note says it is not authoritative and which already carries a known error (lists PSO as an applied thesis method, corrected 2026-08-02). User confirmed both grades directly: **M.Eng. 1.6, B.Eng. 2.4** (German scale, 1.0 best / 4.0 lowest pass). US 4.0 equivalents added via the modified Bavarian formula (`5.0 − G`): **3.4** and **2.6**. ECTS percentile **settled the same day**: B is formally the next 25% after the top 10%, i.e. the **10–35% band, so "top 35%" is the correct phrasing** — the record's original value was right and the briefly-considered "top 30%" was the error. **Never narrow B to a point estimate** (top 30%/25%); it is a band. New `grade_output_rule`: grades stay off the resume and cover letter by default; these fields exist for form questions and credential checks. | `claims.json` (EDU-MENG, EDU-BENG) |
| 2026-08-25 | **Trading-domain evidence was entirely missing from the KB.** `claims.json` had 0 hits for trading, finance or quant, so a genuine capability was unusable in any document. Added **PP-1** (self-built, self-hosted single-user investing/signal platform: US-equity price/fundamental/sentiment ingestion, LLM-assisted sentiment, long-only cross-sectional residual 12-1 momentum book with ATR stop/trail and max 15 names, scheduled scan/backtest pipelines, web dashboard + Telegram alerts) and **PP-2** (Udacity *AI for Trading* nanodegree), plus two `allowed-with-context` skills. Both carry hard `forbidden` lists: **no PnL/Sharpe/return/backtest figure may ever be quoted**, never call PP-1 distributed/production/multi-user, never imply professional quant or finance experience, never present the nanodegree as a quant credential. PP-1 is the one place full-ownership verbs are correct — it is genuinely solo, unlike the employer work governed by Scope Discipline. | `claims.json` (claims PP-1, PP-2; skills "quantitative/systematic trading concepts", "backtesting") |
| 2026-08-02 | **Bosch data types named.** Fab sensor/process data: defect-management records, wafer inspection images, PCM electrical parameters (user-confirmed). Previously the KB named no concrete sensor-data types for Bosch. | `claims.json` BS-2 |
| 2026-07-27 | **Bullet density.** Fixed 1L/2L/3L and character-band rules made bullets uniform and encouraged page filling. | Replaced globally with natural-length, evidence-led bullets; character counts are diagnostic only. |
| 2026-08-25 | **Cadence monoculture across every generated package.** Audit of all 18 packages in `output/` (368 bullets) found **45% used the same "X, Y and Z" triple**; the SBB package hit **85% — 11 of 13 bullets, every Swisscom and Bosch line** — plus two adjacent bullets both opening "Build and…". Not a truth defect: every bullet was accurate and the corpus is clean on AI vocabulary (0 cliché hits in 35 documents) and prose em-dashes (0.08/bullet). It is a *rhythm* tell — uniform three-beat cadence reads as machine-written. PDF metadata is clean (MiKTeX pdfTeX, empty Author/Title, no AI tokens in 69 PDFs, no generator-term leakage). | New `resume_reference.md` §7a + verification step 7; `critical_rules.md` rule 7a; `critique_framework.md` mechanics row now scores cadence (max 1 pt, style only) |
JDs/JD_CitadelSecurities_PlatformEngineer.txt
+59
View File
@@ -0,0 +1,59 @@
SOURCE URL: https://www.citadelsecurities.com/careers/details/platform-engineer/
RETRIEVED: 2026-08-25 via job_scout/.venv Playwright (chromium, headless)
METHOD NOTE: WebFetch returns HTTP 403 on this host; Playwright required.
VERBATIM: yes - full visible posting body, unedited except removal of site
navigation chrome (menu, footer, cookie/legal links).
========================================================================
Platform Engineer
New York, Miami, Zurich
Job Description
Role Overview:
Citadel Securities is seeking an exceptional Platform Engineer to join our Research Platform Engineering team in Miami, Zurich or New York. Our Research Platform teams are at the forefront of designing and developing the distributed systems, data platforms, and research infrastructure that power large-scale simulations, analytics, and model development across the firm. These systems are critical in allowing researchers to transform data into actionable insights and alpha-generating trading strategies.
Opportunities may be available from time to time in any location in which the business is based for suitable candidates. If you are interested in a career with Citadel, please share your details and we will contact you if there is a vacancy available.
Responsibilities:
Design and build distributed platforms and services that support quantitative research, simulations, AI-powered applications, and large-scale analytics workflows
Develop scalable platform services and data systems for ingestion, transformation, storage, and lifecycle management of large datasets
Build and maintain backend services, APIs, and SDKs that improve researcher productivity and self-service capabilities
Improve performance, reliability, and scalability of critical research systems operating in high-throughput environments
Collaborate closely with researchers and engineers to translate research requirements into production-read systems
Qualifications:
Bachelor’s degree in Computer Science or a related field
3+ years of professional software engineering experience
Strong programming skills in Python, Go, or C++, or similar systems-oriented languages
Experience building and operating distributed systems in production environments
Experience designing scalable backend services, APIs, and data-intensive applications
Strong understanding of distributed systems, data-intensive applications, and system design fundamentals
Experience with cloud platforms (AWS, GCP, Azure) and modern infrastructure technologies
Experience with technologies such as Kafka, Kubernetes, Spark, Airflow, distributed databases, or similar systems is a plus
In accordance with applicable law, the base salary range for this role is $175,000 to $350,000.
In addition, the employee who fills this role will be eligible to participate in a discretionary incentive compensation program, as well as a wide array of benefit programs, such as medical and life insurance, retirement and tax-free savings plans, and access to other healthcare programs.
About Citadel Securities
Citadel Securities is a technology-driven, next-generation global market maker. We provide institutional and retail investors with world-class liquidity, competitive pricing and seamless front-to-back execution in a broad array of financial products. Our teams of engineers, traders and researchers harness leading-edge quantitative research and the accelerating power of compute, machine learning and AI to power our analytics and tackle the market’s and our clients’ most critical challenges. Together, we are forging the future of capital markets. For more information, visit citadelsecurities.com.
About Citadel Securities
Citadel Securities is a technology-driven, next-generation global market maker. We provide institutional and retail investors with world-class liquidity, competitive pricing and seamless front-to-back execution in a broad array of financial products. Our teams of engineers, traders and researchers harness leading-edge quantitative research and the accelerating power of compute, machine learning and AI to power our analytics and tackle the market’s and our clients’ most critical challenges. Together, we are forging the future of capital markets. For more information, visit citadelsecurities.com.
Start Application
Citadel Securities
Counterparty Login
Who We Are
What We Do
News & Insights
Careers
Privacy
Terms of Use
Fraud Notice
Disclosures
Copyright © Citadel Enterprise Americas LLC or one of its affiliates. All rights reserved.
Citadel Securities is an equal opportunity employer. We provide all individuals consideration for employment and advancement opportunities without regard to race, religion, color, gender, pregnancy, national origin, age, disability, military or veteran status, sexual orientation, genetic information and any other classification protected by applicable federal, state and local laws.
JDs/JD_SBB_DataEngineer_AssetMgmt.txt
+83
View File
@@ -0,0 +1,83 @@
Data Engineer (m/w/d) - Asset Management Infrastrukturanlagen
Jetzt bewerben
Beweg die Schweiz mit uns.
Data Engineer (m/w/d) - Asset Management Infrastrukturanlagen
Bern / Work Smart, ab sofort oder nach Vereinbarung, 60-100%
Jetzt bewerben
Darauf kannst du dich freuen.
Benefits wie kostenloses GA, vergünstigtes Reisen im Ausland - auch für die Familie - findest du auf unserer Website unter Benefits. 
Das kannst du bewegen.
In dieser spannenden Rolle baust du Datenpipelines auf, betreibst und entwickelst sie weiter auf der Enterprise Analytics Plattform der SBB, die Batch- und Streaming-Ansätze vereint.
Unterschiedlichste Datenquellen integrierst du via APIs/Kafka und ermöglichst damit Analysen, Dashboards sowie vorausschauende Wartung.
Datenprodukte und semantische Modelle gestaltest du auf Snowflake und Power BI.
Du entwickelst RAMSI weiter – von der Webanwendung (Java Spring Boot/Angular) über Reports bis hin zu Dashboards für unsere Fachspezialist:innen.
In enger Zusammenarbeit mit dem Business nimmst du Anforderungen auf und sorgst für nachhaltige Lösungen im Betrieb.
Dein Wissen und deine Erfahrung teilst du im Team und trägst zur fachlichen Weiterentwicklung bei.
Das bringst du mit.
Starke Kommunikationsfähigkeiten und Teamorientierung mit der Bereitschaft, gemeinsam mit unseren Fachspezialist:innen konkrete Lösungen zu erarbeiten und die fachliche Entwicklung anderer Dateningenieur:innen zu fördern.
Agile, selbständige und lösungsorientierte Arbeitsweise mit ausgeprägter Verlässlichkeit und Bereitschaft, Verantwortung für die Qualität und Nachhaltigkeit technischer Lösungen zu übernehmen.
Mehrjährige praktische Erfahrung im Entwurf, Aufbau, der Entwicklung und dem Betrieb cloudbasierter Datenplattformen mit modernem Data Stack (ETL/ELT, Data Warehouses, Streaming, Orchestrierung – z. B. Snowflake, DBT, Helm, Argo Workflows, Kafka).
Fähigkeit, skalierbare und kosteneffiziente Datenarchitekturen zu entwerfen und Trade-offs zwischen Latenz, Konsistenz und Kosten kritisch zu beurteilen.
Erfahrung in Datenmodellierung (Star-/Snowflake-Modelle) und Reporting mit Power BI, fortgeschrittene SQL-Kenntnisse sowie Kenntnisse zur Abfrageleistungsoptimierung.
Software-Engineering-Fähigkeiten in Python, inklusive Tests, CI/CD, Packaging- und Deployment-Best-Practices.
Dein Lohn und deine Benefits.
Wir haben ein faires und fortschrittliches Lohnsystem und setzen uns aktiv für Lohngleichheit ein. Deshalb sind weiterführende und detaillierte Informationen zum Lohnsystem bei der SBB transparent einsehbar. Diese Stelle befindet sich im Anforderungsniveau K, Regionalzulage Stufe 1. Mehr zu Lohn und Benefits.
Damit gelingt der Einstieg.
Dass du mit uns die Schweiz bewegen willst, ist für uns die grösste Motivation. Darum verzichten wir bei dieser Stelle auf ein Motivationsschreiben. Falls du uns deine Motivation dennoch mitteilen möchtest, hast du in unserem Bewerbungstool nach wie vor die Möglichkeit dazu.
Grosses machen ist einfacher per du. Das beginnt bereits bei deiner Bewerbung.
So geht's weiter.
Bewerbung
Virtuelles Kennenlernen mit HR und Führungskraft
Persönliches Kennenlernen
Finaler Entscheid
Noch Fragen?
Zum Bewerbungsprozess
SBB Recruiting Support
Telefon + 41 51 220 20 29
recruiting@sbb.ch
#einfachDu
FAQ zur Bewerbung
Zur Stelle
Job ID 103755
Andri Wienandts, People Leader
Telefon +41 79 364 62 53 
Bewerbungen per E-Mail können leider nicht berücksichtigt werden.
Für diese Stelle nehmen wir keine Dossiers von Personalvermittlern entgegen.
Das bieten wir.
Benefits
Wo immer möglich setzen wir auf flexible Arbeitszeiten und moderne Arbeitszeitmodelle. Darüber hinaus sind wir als Arbeitgeberin auch flexibel, wenn es darum geht, dein Berufsleben mit deinem Privatleben in Einklang zu bringen.
Mehr erfahren
Unterschiede willkommen
Mit über 150 verschiedenen Berufen bieten wir eine beeindruckende Vielfalt an Einstiegs- und Entwicklungsmöglichkeiten. Ebenso vielfältig sind unsere Mitarbeitenden. Darum engagieren wir uns dafür dass alle Mitarbeitenden ihre individuellen Stärken und Denkweisen einbringen können.
Mehr erfahren
Karriere und Entwicklung
Wir fördern unsere Mitarbeitenden entlang ihren persönlichen Bedürfnissen und Ambitionen. Dafür stellen wir ein breites Angebot an fach- und bereichsübergreifenden Weiterbildungsangeboten zur Verfügung. An ausgewählten externen Weiterbildungen, beteiligen wir uns finanziell und zeitlich.
Mehr erfahren
Wir bewegen heute schon Grosses.
Jetzt bewerben
Grosses bewegen.
config.md
+5 -2
View File
@@ -10,7 +10,7 @@
- **Degree suffix:** M.Eng.
- **Email:** dennis@thiessen.io
- **Phone:** +41 795 955 585
- **Location:** Bern, Switzerland
- **Location (contact line):** **Bern, Switzerland** by default. Use **Thun, Switzerland / Thun, Schweiz** for roles in **Thun or its immediate surroundings** - the user lives in Thun, and local residency at the employer's own location is a real screening signal for traditional Swiss employers (user-decided 2026-08-27). This applies to the **contact line only**; a position's own location stays factual, e.g. the Swisscom entry keeps Bern.
- **LinkedIn:** linkedin.com/in/dennis-thiessen
- **Google Scholar:** [leave blank — not applicable]
- **ORCID:** [leave blank — not applicable]
@@ -52,6 +52,7 @@ Verified errors to never re-introduce. Add entries as you catch mistakes.
| Correction | Details |
|-----------|---------|
| Capgemini omitted from all documents | **Standing user preference.** Capgemini Deutschland GmbH (Nov. 2014 - Mai 2015) must **never** appear on a resume, CV or cover letter - only six months, and the user does not want the short-stay impression. The visible chronology starts at **Generali, Mai 2015**. **The resulting Nov. 2014 - Mai 2015 gap is intentional and must never be "fixed" or explained away.** Any tenure claim in a summary line ("over ten years") is measured against the visible record from Mai 2015. Regressed once into the RUAG C5I draft on 2026-08-27 because this rule lived only in memory. |
| Degree name | B.Eng. official name: "Information and Telecommunication Technologies". M.Eng. official name: "Computer Aided Engineering" with focus in Software Design and Software Engineering. Use "Software Design & Engineering" as the focus description on resumes — more recognizable than the official programme name. |
| Education dates | B.Eng. **Oct 2009 – Oct 2012** (start 10/2009, finished 01.10.2012). M.Eng. **Apr 2012 – Oct 2013** (01.04.2012 – 01.10.2013). Programmes overlap by design — do NOT "fix" the overlap. Both at Universität der Bundeswehr München. |
| Swisscom title | Senior: Oct 2023 – Apr 2025. Staff (Engineer IV): Apr 2025 – Present. Use "Staff Data, Analytics & AI Engineer" for current role; note promotion if space allows. |
@@ -114,4 +115,6 @@ These are copied verbatim from your template every time.
- **Local alternative:** 2-page Swiss/DACH profile
- **Cover letter:** Generate only when required or when it adds information not visible in the resume
- **Academic CV:** Not a default target. Generate only for an explicitly academic/research application
- **Output .tex files ONLY** — user compiles locally
- **The `.tex` is the deliverable and the source of truth** - never hand-edit a PDF, and always regenerate from the source
- **Compile the PDF into the application's own `output/` folder** so the user can review and approve it. A `.tex` alone is not reviewable for him. Never leave the only PDF in a temp directory, and never leave a stale pre-edit PDF next to the deliverable (user-reported 2026-08-27).
- **If a submission-named copy exists** (e.g. `Dennis_Thiessen_Lebenslauf.pdf`), **refresh it on every recompile.** It is a copy, not a build target, so it silently goes stale under exactly the filename that gets attached to an application.
job_scout/ROSTER_REVIEW_2026-08-18.md
+205
View File
@@ -0,0 +1,205 @@
# Scout roster review — 2026-08-18
> **Status: implemented 2026-08-18.** Sections 1, 2 and 5 have been acted on; see
> "What was actually done" at the bottom, which also records two findings in this
> document that turned out to be **wrong** once the fixes were attempted.
Originally analysis only. Every claim below was verified by probing the
live endpoint; unverified candidates are listed separately as *probe-worthy*, not as recommendations.
Method: aggregated the scan-stats tables from the four most recent full runs (2026-07-14, 07-27,
08-07, 08-18), cross-referenced `state/decisions.json` (253 decisions), and live-probed candidate
ATS endpoints.
---
## 1. Three scrapers are silently broken
Same failure class as the dbt Labs 404 fixed earlier today, but worse: these return HTTP 200 and
plausible-looking rows, so the report shows "0 matches" and nothing looks wrong. **A company that
scrapes garbage is indistinguishable from a company with no openings.**
| Company | What it actually returns | Verified |
|---|---|---|
| **Roche** | Shanghai, Kyiv, Petaling Jaya, Bogota, Berlin, Mannheim, Penzberg. **Zero Swiss roles** despite Basel HQ — the `locationsearch=Switzerland` URL param is being ignored and an unfiltered default page is scraped. | live scrape, 10/10 rows non-CH |
| **Meta** | **1 row**: "Research Scientist Intern, AI/ML, Core Ads Growth (PhD)". Meta Zurich is a real engineering office. The `offices[0]=Zurich` filter or the lazy-load scroll is failing. | live scrape, 4 jobs across 4 runs |
| **Apple** | 21 rows, **all** internships, "Where we're hiring", and "US - Specialist: Seasonal, Part-time". Zero actual engineering reqs. | live scrape, all 21 rows |
**This invalidates a conclusion in the yield data.** Apple shows 21 decisions, all `skip`, 0 signal —
that reads as "Apple is a bad fit" but it is really "the scraper only ever saw the internships page."
Don't drop Apple on that evidence; fix the scrape first, then judge.
## 2. Seven boards have no title filter at all
`_title_filter` is absent on **BKW, SBB, BFH, PostFinance, Swissgrid, Novartis, Palantir**. This is
why reports drown in *Mitarbeiter/in Hausdienst*, *Financial Accountant*, *Studiengangsleiter Master
Physiotherapie*, *SAP Consultant Treasury*.
| Company | Eligible → match ≥2 (4 runs) | Reading |
|---|---|---|
| BKW | 61 / 75 scraped | 81% of the entire board "matches" |
| SBB | 27 / 27 | 100% — filter is doing nothing |
| BFH | 29 / 48 | 60% |
| Novartis | 105 eligible → 21 match | 23 skips vs 3 shortlists |
**These need a title filter, not removal.** BKW/SBB/BFH/PostFinance/Swissgrid are the Bern/Thun
WLB-exception tier where below-bar comp is deliberately accepted — the lane is intentional, the
noise is not. RUAG already carries a 29-term filter and behaves (13 eligible → 7 match); it is the
model to copy.
## 3. The yield metric in the decisions log is confounded
Raw decision counts reward boards with *no* title filter — more rows surfaced means more `skip`
rows recorded. Normalising to signal (`shortlist`+`applied`+`maybe`) per eligible role inverts the
ranking: **Snowflake** (5 decisions → 1 applied, 1 shortlist, 1 maybe off 14 eligible) is a far
better board than **Novartis** (31 decisions → 3 shortlists off 105 eligible). Worth keeping in
mind before reading the decisions log as a quality signal.
## 4. Dead lanes
- **MET Group + Louis Dreyfus: 1,464 roles scraped across 4 runs → 0 eligible, 0 matches. Ever.**
Both are title-filtered correctly; they simply have no CH/remote-CH engineering roles. LDC is
Geneva/Rotterdam-anchored (French). The commodity-trading lane is configured and producing nothing.
- **Norway (Telenor, Equinor, NATO JWC Stavanger)** — three of 34 companies on a lane your own notes
call mostly closed: comp below the 180k bar, Norwegian at A2, no clearance. Equinor just came back
rejected at 79 days. Consolidation candidate, though NATO JWC has an open application pending.
---
## 5. Verified adds — both confirmed live
### Amazon / AWS — the single biggest gap
`https://www.amazon.jobs/en/search.json?normalized_country_code[]=CHE` → clean public JSON, **32 CH roles**.
Among them, live right now:
- **Senior Forward Deployed Engineer, AWS Forward Deployed Engineering — Zurich**
- Sr. Delivery Consultant – AI/ML, Professional Services — Zurich
- Sr Specialist SA – GenAI, Specialists Team Germany/Switzerland — Zurich
- Senior Security Assurance Solutions Architect, AWS Security Assurance — Zurich
Why this matters more than any other candidate: **AWS is the primary cloud in the evidence base**
(SW-1 migration, the Swisscom data products). `claims.json:274` marks GCP `evidence: unverified,
output: forbidden` — which is exactly what hard-gated the two Google GenAI FDE reqs this morning.
An AWS-native FDE req in Zurich is the same lane without the cloud gap. Clears the comp bar.
No adapter work needed beyond a thin JSON fetcher.
### Axpo — the energy-trading lane, finally live
`https://careers.axpo.com/jobs.json` — Teamtailor on a custom domain. **400 roles, 98 tech-ish.**
Note the feed is JSON Feed format keyed `items`, not `jobs`, so the existing `fetch_teamtailor`
needs a small tweak (and the demo-board fingerprint guard still applies).
Live right now: **"Forward Deployed AI Engineer (f/m/d)"**, plus Senior Application Manager/Solution
Architect, Quantitative Modeller, Junior/Financial Data Analyst, Leiter/in IT Division Hydroenergie.
Memory (`user_role_targeting_energy_trading`) names Axpo and Alpiq explicitly as targets — data/
platform engineering *inside* a trading shop, not ETRM or quant. Only MET was ever configured, and
MET has produced nothing. Axpo is German-speaking (Baden/Zurich), which fits the DE/EN profile that
rules out the Geneva traders.
## 6. Verified negative — do not add
- **Coinbase proper.** Greenhouse board `coinbase` is live (169 jobs) but has exactly **2 EMEA-remote
roles, both non-engineering** (Business Controller, Threat Assessment Manager). The Ventures board
removed earlier was portfolio-only, so Coinbase was never really covered — but covering it properly
gains nothing. Crypto lane stays Kraken + Bitcoin Suisse.
- **Alpiq, Glencore, Swiss Re, Sunrise** — no SmartRecruiters, Greenhouse, Ashby or Lever board found.
Careers sites are bespoke/JS-rendered with no discoverable JSON API in a network trace. Not worth
Playwright maintenance on spec.
- **The remote-EU data-infra tier as a category** (ClickHouse, MongoDB, Redis, Temporal, Airbyte,
Starburst…). Skipping this deliberately: Grafana is already on the roster and memory flags it as
**below-bar geo-fenced comp**. These firms band EU-remote salaries the same way, so the whole tier
fails the 180k bar regardless of role fit. Elastic/Confluent/Grafana already cover the archetype.
## 7. Probe-worthy — not yet verified, do not add on my word
- **CERN** (`careers.cern/jobs`) — has dedicated "Data Science, AI & Analytics" and "Information
Technologies" fields of work. English-working international organisation in CH, i.e. the same lane
as the BIS Basel application. Needs a scrape-mechanics check. Best remaining unverified candidate.
- **Swiss Post** (`career.post.ch`) — Bern-headquartered, has an "Informatik und Digital Services"
category, fits the Bern WLB tier alongside PostFinance (already scraped, and a Swiss Post company).
My probe timed out on `networkidle`; needs a retry with a laxer wait.
- **NATO NCIA** (`ncia.nato.int/careers.html`) — the current `nato` adapter returns 4 reqs, all JWC
Stavanger. NCIA is NATO's actual technology arm (Brussels/The Hague/Mons) and is a separate
pipeline. Worth checking whether it is separately scrapable — it would be the one way to keep a
NATO lane without the Norway constraints.
- **SIX Group** — fingerprinted as SuccessFactors (`career_company=sixgroupse`). Zurich financial-
market infrastructure. Endpoint pattern is known; content unverified.
---
## Suggested order of work
1. **Fix Roche / Meta / Apple** — restores three companies that currently contribute nothing while
appearing healthy. Highest value per unit effort, and it's a correctness bug, not a preference.
2. **Add Amazon/AWS** — trivial adapter, an on-profile Zurich FDE req live today.
3. **Add Axpo** — small `fetch_teamtailor` tweak for the `items` key; opens the energy lane.
4. **Add title filters** to BKW, SBB, BFH, PostFinance, Swissgrid, Novartis, Palantir.
5. **Decide on MET/LDC and the Norway three** — a judgment call, not a bug.
6. Probe CERN, Swiss Post, NATO NCIA, SIX.
---
# What was actually done (2026-08-18)
Implementing the recommendations disproved two of them. Both corrections are recorded here
rather than quietly edited above, because the original claims were used to justify the work.
## Correction 1 — Meta is NOT broken
Section 1 listed Meta as returning garbage. It does not. `metacareers.com` itself reports
**"1 Items"** for the Zurich office filter — the board is genuinely near-empty and the scraper
reports it accurately. No change made beyond a comment recording the verification, so it is
not "fixed" again later. Roche and Apple were real bugs; Meta was not.
## Correction 2 — inclusion filters were the wrong instrument
Section 2 recommended adding `_title_filter` allowlists to seven boards. That was implemented
and then **reversed**, on the objection that an allowlist *fails closed*: a strong-fit role
with a title nobody anticipated is dropped at fetch time, never scored, and appears in no
report and no JSON dump. The final call should sit with the scorer and the reviewer, not a
keyword gate.
Filtering is now two explicit modes:
| Mode | Behaviour | Where |
|---|---|---|
| `_title_exclude` (**default**) | Fails **open** — drops only unambiguous non-tech titles (`hausdienst`, `physiotherapie`, `violine`, `lehrstelle`, `legal counsel`…). Everything else is scored. | 14 boards |
| `_title_filter` | Fails **closed** — allowlist, used only where volume makes full scoring impractical (>~200 roles) | Databricks, Snowflake, Datadog, Elastic, Fivetran, Louis Dreyfus |
`NOISE_TITLE_EXCLUDE` is one shared, board-agnostic list. Terms that could plausibly attach to
a technical role (analyst, manager, specialist, consultant, architect, lead) are deliberately
excluded from it.
Also corrected: BKW's 61/75 match rate was blamed on the missing filter. The real cause is
`_score_floor: 2`, set deliberately because the English keyword scorer cannot read German
titles. The floor was left in place.
## Fixes shipped
- **Roche** — new `fetch_phenom` adapter (Phenom `refineSearch`). Was 40 scraped / 0 Swiss;
now **133 scraped, 88 CH-eligible, 25 matches**.
- **Apple** — removed `default_location: "Switzerland"`, which was relabelling US "Various
Locations" postings as Swiss. Now honestly reports **0 CH-eligible**.
- **Amazon / AWS** — new `fetch_amazon` adapter. **32 CH roles, 29 eligible, 13 matches**,
including the Zurich AWS FDE req and a Senior ProServe Cloud Architect in **Bern**.
- **Axpo** — `fetch_teamtailor` extended for custom domains and pagination. 461 roles.
Locations now come from the feed's schema.org `jobLocation`, with ISO alpha-2 expanded to
full country names — without that, `"Burgdorf, CH"` failed the CH keyword match, and a
forced `default_location` would have marked ~380 Madrid/Milan/Warsaw roles as Swiss (the
Apple bug again). Telenor gained real locations from the same change.
- **Filter modes** as described above, plus `artificial intelligence`, `data scientist`,
`data-driven`, `ai engineer`, `ai platform`, `ai architect`, `ai-systems`, `data science`
added to the shared allowlist. Bare `"ai"` deliberately omitted — it substring-matches
*Maintenance*, *Training*, *Chair*.
- **Palantir left unfiltered** — its entry already documents why (its target titles, e.g.
"Deployment Strategist", are not in the allowlist). The original recommendation to filter it
would have hidden exactly the roles worth seeing.
## Still open (deliberately not done)
- **Section 4 judgment calls**: MET Group + Louis Dreyfus, and the three Norway companies.
These are preference decisions, not bugs.
- **Section 7 probes**: CERN, Swiss Post, NATO NCIA, SIX Group.
job_scout/scout.py
+439 -62
View File
@@ -138,6 +138,9 @@ NEGATIVE_KEYWORDS = {
# post mostly non-tech roles). Only keep titles containing one of these specific role
# phrases — kept tight so "Sales Engineer"/"Staff Accountant"/"Data Privacy Counsel"
# don't leak in. Matched as case-insensitive substrings against the title only.
# Inclusion allowlist. Applied ONLY to boards too large to score in full (>~200 roles:
# Databricks, Snowflake, Datadog, Elastic, Fivetran, Louis Dreyfus, Palantir). Everywhere else
# use NOISE_TITLE_EXCLUDE — see the design note there for why fail-open is the default.
ENG_TITLE_FILTER = [
"data engineer", "data engineering", "data platform", "platform engineer",
"data infrastructure", "data architect", "analytics engineer",
@@ -149,8 +152,52 @@ ENG_TITLE_FILTER = [
# "resident" alone catches Resident Solutions Architect/Engineer without opening the gate to
# all pre-sales SAs (the overscoring trap); "customer engineer" is Google's field-eng term.
"forward deployed", "forward-deployed", "field engineer", "resident", "customer engineer",
# Added 2026-08-18 after the CH boards were title-filtered: without these, real near-misses
# were dropped (BFH "Wissenschaftliche Mitarbeit Data-Driven Government", Novartis "Director
# & Group Head (AI-Systems & Scale)"). Deliberately NOT adding bare "ai" — it substring-matches
# Maintenance/Training/Chair and floods every board.
"artificial intelligence", "ai-systems", "ai platform", "ai engineer", "ai architect",
"data scientist", "data-driven", "data science",
]
# Generic non-tech title exclusions, shared by every board that uses exclusion-mode filtering.
#
# DESIGN NOTE (2026-08-18). Boards are filtered one of two ways:
# * inclusion (`_title_filter`) — an allowlist, used ONLY on boards too large to score in
# full (Databricks ~800, OpenAI ~730, Palantir ~300). It fails CLOSED: a great-fit role
# with a title nobody anticipated is dropped at fetch time and never reaches the scorer,
# so it appears in no report and in no JSON dump. That is an acceptable trade only where
# volume forces it.
# * exclusion (`_title_exclude`) — this list. It fails OPEN: everything survives unless it
# is *clearly* not an engineering role, and the score + the weak/noise bucket decide what
# surfaces. Prefer this. The final call belongs to the reviewer, not to a keyword gate.
#
# Keep these terms unambiguous. Anything that could plausibly attach to a technical role
# (analyst, manager, specialist, consultant, architect, lead) must NOT go in here.
NOISE_TITLE_EXCLUDE = [
# Facilities / retail / hospitality / admin
"hausdienst", "reinigung", "empfang", "hauswart", "chef de partie", "koch", "küche",
"restaurant", "catering", "fahrer", "logistik mitarbeiter", "lagerist", "verkauf",
"verkäufer", "retail", "barista", "security guard", "sicherheitsdienst",
# Care / health / teaching-of-non-tech
"physiotherapie", "pflege", "pflegefach", "hebamme", "ergotherapie", "psychologie",
"medical representative", "nurse", "arzt", "ärztin", "dentist",
# Arts / music / sport
"violine", "klavier", "musik", "dozierende*r violine", "sport",
# Back-office
"fundraising", "buchhaltung", "accountant", "accounting", "payroll", "steuer",
"recruiter", "talent acquisition", "human resources", "personalwesen",
"legal counsel", "rechtsanwalt", "rechtsreferendar", "notar", "jurist",
"kommunikation", "public relations", "übersetzer", "translator",
# Early-career / non-role listings
"lehrstelle", "praktikum", "praktikant", "internship", "intern ", "trainee",
"apprenti", "ausbildung", "schnupper", "where we're hiring", "talent community",
# Non-software engineering trades
"elektroplaner", "elektroinstallat", "sanitär", "heizung", "maler", "schreiner",
"hochspannung", "wasserbau", "strassenbau", "holzbau", "bauingenieur", "bauleiter",
]
# id, display, adapter, adapter_args
COMPANIES = [
("nvidia", "NVIDIA", "workday", {
@@ -167,6 +214,7 @@ COMPANIES = [
"tenant": "novartis",
"site": "Novartis_Careers",
"search_text": "Switzerland",
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# PCSX (Eightfold) — Microsoft has a public position search endpoint
("microsoft", "Microsoft", "pcsx", {
@@ -177,22 +225,56 @@ COMPANIES = [
# Dropped: ClickHouse (Glassdoor 3.3, 36% recommend, toxic-culture flag — 2026-05).
# Dropped: HashiCorp — acquired by IBM (closed 2025); greenhouse/ashby/lever boards all 404,
# roles folded into IBM's careers (no clean public ATS API). 2026-06-06.
("confluent", "Confluent", "ashby", {"slug": "confluent", "_title_filter": ENG_TITLE_FILTER}),
("gitlab", "GitLab", "greenhouse", {"board": "gitlab", "_title_filter": ENG_TITLE_FILTER}),
("grafana", "Grafana Labs","greenhouse",{"board": "grafanalabs", "_title_filter": ENG_TITLE_FILTER}),
("confluent", "Confluent", "ashby", {"slug": "confluent", "_title_exclude": NOISE_TITLE_EXCLUDE}),
("gitlab", "GitLab", "greenhouse", {"board": "gitlab", "_title_exclude": NOISE_TITLE_EXCLUDE}),
("grafana", "Grafana Labs","greenhouse",{"board": "grafanalabs", "_title_exclude": NOISE_TITLE_EXCLUDE}),
# Added 2026-06-06 (Tier A/B data-infra). Databricks/Snowflake/Datadog have Zürich offices
# (Swiss-scale comp, clears bar); Elastic/dbt Labs are remote-EU (verify CH-equiv comp —
# (Swiss-scale comp, clears bar); Elastic/Fivetran are remote-EU (verify CH-equiv comp —
# may be geo-banded below 180k, like Grafana). All title-filtered (boards are 160-760 roles).
("databricks","Databricks","greenhouse", {"board": "databricks", "_title_filter": ENG_TITLE_FILTER}), # Zürich SWE + remote-EU; DK via default location policy
("snowflake", "Snowflake", "ashby", {"slug": "snowflake", "_title_filter": ENG_TITLE_FILTER}), # Zürich "Observe" observability SWE roles
("datadog", "Datadog", "greenhouse", {"board": "datadog", "_title_filter": ENG_TITLE_FILTER}), # Zürich branch + remote-EU
("elastic", "Elastic", "greenhouse", {"board": "elastic", "_title_filter": ENG_TITLE_FILTER}), # remote-first; ELK = his stack
("dbtlabs", "dbt Labs", "greenhouse", {"board": "dbtlabsinc", "_title_filter": ENG_TITLE_FILTER}), # remote-EU; analytics-eng
# dbt Labs -> Fivetran: the "dbtlabsinc" board 404'd from 2026-08-18. getdbt.com/careers now
# redirects to a page serving Fivetran's Greenhouse listings (gh_jid links) — the two merged, so
# the dbt roles live on board "fivetran". EMEA presence is Dublin/London/Serbia, no CH office, so
# most roles fall out on the location policy; kept for remote-EU analytics-eng reqs. 2026-08-18.
("fivetran", "Fivetran (ex-dbt Labs)", "greenhouse", {"board": "fivetran", "_title_filter": ENG_TITLE_FILTER}),
# --- Energy / commodity trading (SmartRecruiters; title-filtered to tech roles) ---
# Dropped: Vitol (Glassdoor 3.5, 55% recommend, grueling-hours/toxic flag — 2026-05).
# Dropped: Sygnum (Glassdoor 3.4, 51% recommend, comp 2.3/5 — below 180k bar — 2026-05).
("metgroup", "MET Group", "smartrecruiters", {"company": "METGroup", "_title_filter": ENG_TITLE_FILTER}),
("ldc", "Louis Dreyfus","smartrecruiters",{"company": "LouisDreyfusCompany", "_title_filter": ENG_TITLE_FILTER}),
# Axpo (Baden/Zurich) — Teamtailor on a custom domain; no <slug>.teamtailor.com host
# resolves, so this uses base_url. Added 2026-08-18: memory names Axpo/Alpiq as the
# energy-trading targets but only MET was ever configured, and MET+LDC have produced
# 0 CH-eligible roles across 1,464 scrapes. ~400 roles, German-speaking region (which is
# what rules out the Geneva traders), incl. a "Forward Deployed AI Engineer" at add time.
("axpo", "Axpo", "teamtailor", {
"base_url": "https://careers.axpo.com",
"default_location": "Switzerland",
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# CERN (Geneva) — probed 2026-08-18: SmartRecruiters board "CERN", 66 live postings, all
# Geneva. Real engineering exists (Deep Learning Developer, ML Engineer for Next Generation
# Triggers, Full-stack SWE, Data Storage R&D Engineer, Software Engineer) and it is an
# English-working international organisation in CH — the BIS Basel lane.
# TWO STANDING CAVEATS, do not forget them when scoring a CERN role:
# 1. Geneva is ~1h50 each way from Bern — not viable for a 2-3 day hybrid without
# relocation, which the user has ruled out. Practical constraints will score low.
# 2. CERN pay grades sit below the CHF 180k bar.
# Added anyway because the adapter is free (already implemented) and a fully-remote or
# exceptional CERN req would otherwise never surface.
("cern", "CERN (Geneva)", "smartrecruiters", {
"company": "CERN",
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
("metgroup", "MET Group", "smartrecruiters", {"company": "METGroup", "_title_exclude": NOISE_TITLE_EXCLUDE}),
# Dropped: Louis Dreyfus — 2026-08-18. 300 roles, every one Brazil/India/Bulgaria/Asia
# operations (its Data Modeler reqs are in Bangalore). Geneva HQ but zero Swiss
# postings in any run. MET Group is retained by contrast: once exclusion-mode
# filtering replaced the allowlist, MET showed real Baar/Zug roles (Wholesale
# Specialist, Lead Trade Process Analyst, LNG Financial Trader) — its earlier
# "0 eligible across 1,464 scrapes" was a filter artifact, not an empty board.
# No data/platform eng reqs there yet, but it is the live CH commodity-trading lane.
# Equinor (Workday) — Norway energy major; lived/worked in NO before. Outlier location
# policy: Norway only (not CH / Europe-remote). Small board (~15); no title filter.
("equinor", "Equinor", "workday", {
@@ -207,8 +289,28 @@ COMPANIES = [
# Telenor Norge (Teamtailor) — Nordic telco; Swisscom is the direct domain analogue
# and Dennis worked in Norway before (Vizrt, Bergen). NOTE: the old Workday tenant
# telenorgroup.wd3.myworkdayjobs.com is dead (HTTP 422 on /wday/cxs, 2026-07-29) —
# Telenor migrated to Teamtailor. Slug telenorgroup is a near-empty group-HQ board;
# telenornorway is the real Norway board (redirects to careers.telenor.no).
# Telenor migrated to Teamtailor. telenornorway is the real Norway board (redirects
# to careers.telenor.no).
# TRAP — slug telenorgroup is NOT a "near-empty group-HQ board" as this comment used
# to claim (corrected 2026-08-27): it is a claimed board titled "Sandbox: Telenor
# Shared Services" serving 18 phantom listings whose newest is 2024-11-11, including
# plausible-looking "Lead Software Engineer" / "Frontend Developer - Oslo" roles. None
# of its titles hit the DEMO_TITLES fingerprint in fetch_teamtailor(), so that guard
# would NOT catch it — do not widen DEMO_TITLES to cover this. That guard is for
# *unclaimed* slugs served Teamtailor's demo seed set; this is a claimed board serving
# years-stale data, a different failure mode (a newest-item age threshold is the fix,
# if one is ever wanted). Never point this entry at telenorgroup.
# Sibling Telenor boards probed 2026-08-27, all rejected, do not re-probe:
# telenorcyberdefence (3 roles, newest 2025-10-06 — dormant), telenordenmark (4 roles:
# retail sales, praktikant, legal student — zero engineering, and it would run under
# the default CH/DK policy rather than norway_only, so it is pure noise),
# telenoramp / telenormaritime / telenorsatellite / telenorsverige / telenorkystradio /
# telenorbusiness (all HTTP 404 — no such board).
# THROUGHPUT (measured 2026-08-27): the live board carries 4-5 roles at a time and is
# mostly Norwegian-language ops/security. Exactly one relevant role has surfaced —
# Platform Engineer, Core Automation (job 7829423, posted 2026-06-01, score 5), first
# reported 2026-08-18 and HTTP 410 Gone by 2026-08-27. The feed did NOT silently drop
# it; it genuinely closed. Expect roughly one tailorable role per quarter here.
# CAVEAT: most Telenor Norge postings require Norwegian ("norsk og engelsk") and are
# Oslo/Trondheim onsite-hybrid — verify both per role before tailoring.
# Norwegian-language titles/descriptions score ~0 under the English eng keyword
@@ -240,19 +342,12 @@ COMPANIES = [
"url": "https://www.bis.org/doclist/vacancies.rss",
"default_location": "Basel, Switzerland",
}),
# Coinbase Ventures web3 talent network (Getro collection 1625). Aggregates roles
# across portfolio companies (Notion, Ashby, VALR, World, ...), NOT Coinbase itself —
# see fetch_getro. CH-filtered + eng title-filtered to stay relevant.
("coinbase_ventures", "Coinbase Ventures (web3)", "getro", {
"collection": 1625,
"locations": ["Switzerland"],
"job_functions": ["Software Engineering", "IT", "Data Science"],
# User preference 2026-07-28: recruiting-software domain is not a target.
# Exclude the underlying employer here without hiding unrelated companies
# carried by the Coinbase Ventures portfolio board.
"_exclude_orgs": ["Ashby"],
"_title_filter": ENG_TITLE_FILTER,
}),
# Dropped: Coinbase Ventures web3 talent network (Getro collection 1625) — 2026-08-18.
# Never carried Coinbase's own roles, only portfolio companies. Across ~3 months
# it surfaced 3 distinct CH-eligible roles, ALL from Ashby (recruiting software,
# ruled out 2026-07-28) and all decided "skip". With Ashby excluded the board has
# returned 0 jobs on every run since, reported as a false error each time.
# fetch_getro is retained for a future VC talent network worth tracking.
# Bitcoin Suisse (Zug) uses the onlyfy.jobs ATS. No title filter — small crypto
# firm, only a handful of CH roles; let scoring rank them (CH filter does the rest).
("bitcoin_suisse", "Bitcoin Suisse", "onlyfy", {"slug": "bitcoin-suisse"}),
@@ -280,6 +375,20 @@ COMPANIES = [
"page_param_start": 2,
"max_pages": 6,
}),
# Amazon + AWS share one board. Added 2026-08-18: AWS is the cloud the evidence base
# actually supports (Swisscom migration + data products), whereas claims.json marks GCP
# output-forbidden — so an AWS-native Zurich FDE req is the Google FDE lane without the
# cloud gap. 32 CH roles at time of adding, incl. "Senior Forward Deployed Engineer, AWS
# Forward Deployed Engineering" (Zurich).
# AWS names its delivery/field org differently from every other board, so the shared
# filter caught only 1 of 32 CH roles. These four terms are added narrowly and ONLY here:
# "solutions architect" is the classic overscoring trap (see job_scout_overscoring_findings),
# but on the AWS board it is the delivery-side title attached to migration/data work, which
# is exactly SW-1. Read the JD before trusting the score on any of these.
("amazon", "Amazon / AWS", "amazon", {
"countries": ["CHE"],
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
("apple", "Apple", "playwright", {
"url": "https://jobs.apple.com/en-us/search?location=switzerland-CHE",
"wait_for": "a[href*='/en-us/details/']",
@@ -287,9 +396,19 @@ COMPANIES = [
"title_attr": "text",
"link_attr": "href",
"url_prefix": "https://jobs.apple.com",
"default_location": "Switzerland",
# NO default_location. Apple's own postLocation-CHE filter leaks global reqs whose
# postLocationId is postLocation-USA ("Various Locations within United States"), and
# a forced "Switzerland" default relabelled every one of them as CH-eligible — 84
# phantom CH rows over 4 runs, all internships/retail, which then read as "Apple is a
# bad fit" in the decision log. Read the real location off the card instead; when
# Apple has no Swiss reqs the honest answer is 0. Verified 2026-08-18.
"use_inner_text_as_blob": True,
"scroll_count": 5,
}),
# Meta job links are /profile/job_details/<id>; title + location are in the link text.
# NOT broken despite returning ~1 role: verified 2026-08-18 that metacareers itself
# reports "1 Items" for the Zurich office filter. The board is genuinely near-empty;
# don't "fix" this scraper without first checking the live item count on the page.
("meta", "Meta", "playwright", {
"url": "https://www.metacareers.com/jobs?offices[0]=Zurich%2C%20Switzerland",
"wait_for": "a[href*='/profile/job_details/']",
@@ -301,22 +420,18 @@ COMPANIES = [
"scroll_count": 5,
"use_inner_text_as_blob": True,
}),
# PhenomPeople pattern (Roche) uses li.jobs-list-item.
# Card inner text is structured like: "<title> | Location | <city, country> | Category | ..."
# We extract title from first line, full text becomes the "description" so our location
# filter still sees Switzerland mentions.
("roche", "Roche", "playwright", {
"url": "https://careers.roche.com/global/en/search-results?keywords=&locationsearch=Switzerland",
"wait_for": "li.jobs-list-item, a.au-target",
"card": "li.jobs-list-item:not(:has-text('Saved jobs'))",
"title_attr": "text",
"link_sel": "a[href]",
"link_attr": "href",
"url_prefix": "https://careers.roche.com",
"default_location": "",
"cookie_accept": ["#onetrust-accept-btn-handler", "button:has-text('Accept All Cookies')"],
"scroll_count": 6,
"use_inner_text_as_blob": True,
# Roche: was a playwright scrape of the search-results page until 2026-08-18. That page's
# ?locationsearch=Switzerland filter silently fails — the page state reports "no-results"
# and the scrape harvested *recommendation-widget* cards instead, returning Shanghai,
# Kyiv, Bogota and Mannheim and zero Swiss roles for months while looking healthy. The
# underlying Phenom refineSearch endpoint filters correctly (134 CH roles), so hit it
# directly. Title-filtered: Roche CH is overwhelmingly lab/pharma/apprenticeship reqs.
("roche", "Roche", "phenom", {
"url": "https://careers.roche.com/widgets",
"ref_num": "ROCHGLOBAL",
"page_id": "page11-ds",
"country_facet": "Switzerland",
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# Cisco (PhenomPeople, new careers.cisco.com domain). Keyword search surfaces CH roles.
("cisco", "Cisco", "playwright", {
@@ -356,6 +471,7 @@ COMPANIES = [
"field_url": "descriptionUrl", "field_date": "onlineSince",
"loc_suffix": " Switzerland",
"desc_keys": ["department", "typeOfEmployment", "entryLevel"],
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# RUAG (Thun/Bern/Emmen). Jobs render on the portal as anchors to jobs.ruag.ch; the first
# line of each anchor is the title. All sites are Swiss, so default_location=Switzerland
@@ -373,7 +489,7 @@ COMPANIES = [
"scroll_count": 1,
"page_param": "page",
"max_pages": 10,
"_title_filter": ENG_TITLE_FILTER,
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# SBB (company.sbb.ch — the correct host; company-jobs.sbb.ch was wrong). AEM job filter
# served as a flat JSON list; the fetch_sbb adapter replicates the user's IT + Bern-region
@@ -383,12 +499,14 @@ COMPANIES = [
"topic": "IT / Telekommunikation",
"region": "Bern Mittelland",
"_score_floor": 2,
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# BKW Group (jobs.bkw.com — the real ATS host). PMS structured-data API; ~600 roles
# group-wide, so fetch_bkw keeps only Berufsfeld categories Informatik/Trading/Finanzen
# (IT/data + energy-trading, incl. the flagged Energiehandel roles). German/generic
# titles, so _score_floor keeps the pre-filtered set visible.
("bkw", "BKW (Bern)", "bkw", {"_score_floor": 2}),
("bkw", "BKW (Bern)", "bkw", {"_score_floor": 2,
"_title_exclude": NOISE_TITLE_EXCLUDE}),
# PostFinance (Bern). The careers site renders a small, client-side paginated board;
# scrape all pages through its stable next-page control. No title filter: the board is
# low-volume, and the scorer keeps unrelated banking/customer-service roles out of the
@@ -403,6 +521,7 @@ COMPANIES = [
"use_inner_text_as_blob": True,
"next_button": "#pfch-pagination-next",
"max_pages": 10,
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
# BFH (Bern University of Applied Sciences). Re-added 2026-07-14: the jobs.bfh.ch domain
# itself is a broken/stub SPA shell (renders "Career Center project template", nothing
@@ -422,6 +541,7 @@ COMPANIES = [
"link_sel": "a",
"default_location": "Switzerland",
"_score_floor": 2,
"_title_exclude": NOISE_TITLE_EXCLUDE,
}),
]
@@ -439,6 +559,12 @@ MANUAL_CHECK = [
# (then client-filter requisitionList[].PrimaryLocation for Switzerland/Zürich)
("Oracle", "ORC SPA resists scraping; REST endpoint known but needs CH geographyId (see code comment). Check Switzerland tech roles manually.",
"https://careers.oracle.com/en/sites/jobsearch/jobs?location=Switzerland"),
# NATO NCIA (Brussels/The Hague/Mons) — probed 2026-08-18. NCIA is NATO's technology arm and
# a SEPARATE pipeline from the `nato` adapter above, which only returns JWC Stavanger reqs
# (Taleo). nato.int sits behind Cloudflare and returns "Sorry, you have been blocked" to a
# headless browser, so there is no automated path. Manual only.
("NATO NCIA", "Cloudflare-blocked to headless browsers. NCIA is the tech arm and is separate from the JWC Stavanger reqs the `nato` adapter returns — check it manually for AI/data roles.",
"https://www.ncia.nato.int/careers.html"),
]
@@ -647,34 +773,80 @@ def fetch_rss(args):
return jobs
# schema.org JobPosting addresses carry ISO alpha-2 country codes; the location policy
# keyword lists are full names. Only the countries the policies actually test for.
_ISO2_COUNTRY = {
"CH": "Switzerland", "NO": "Norway", "DK": "Denmark", "DE": "Germany",
"AT": "Austria", "FR": "France", "IT": "Italy", "ES": "Spain", "PL": "Poland",
"NL": "Netherlands", "BE": "Belgium", "SE": "Sweden", "FI": "Finland",
"GB": "United Kingdom", "UK": "United Kingdom", "IE": "Ireland", "US": "United States",
"PT": "Portugal", "CZ": "Czechia", "RO": "Romania", "RS": "Serbia", "TR": "Turkey",
}
def fetch_teamtailor(args):
"""Teamtailor public JSON Feed (`https://<slug>.teamtailor.com/jobs.json`).
Used by Telenor. The public /jobs HTML page paginates and under-reports; the feed
returns the full board in one call, so prefer it. The feed carries no location
field, so default_location is required (Teamtailor boards are per-country anyway).
Used by Telenor (`slug`) and Axpo (`base_url`, a Teamtailor board on a custom domain —
careers.axpo.com — which no <slug>.teamtailor.com host resolves to). The public /jobs
HTML page paginates and under-reports; the feed is authoritative. The feed carries no
location field, so default_location is required (Teamtailor boards are per-country).
Descriptions are frequently Norwegian — see the _score_floor note on the company.
The feed caps at 100 items and links the next page via `next_url`; follow it, or large
boards silently truncate (Axpo is ~400 roles, i.e. 4 pages).
TRAP: an unclaimed slug returns Teamtailor's *demo* board (HTTP 200, plausible JSON)
instead of 404. Verified 2026-07-29: `ksat` and `akerbp` both served the seed set
below though neither company uses Teamtailor. Filtered here so phantom roles never
reach a report — if a real board is ever dropped by this, widen the fingerprint."""
DEMO_TITLES = {"sales development manager", "team lead - csm", "social media manager",
"customer success manager", "key account manager", "ux designer"}
if args.get("base_url"):
url = args["base_url"].rstrip("/") + "/jobs.json"
else:
url = f"https://{args['slug']}.teamtailor.com/jobs.json"
req = urllib.request.Request(url, headers={"User-Agent": USER_AGENT, "Accept": "application/json"})
jobs = []
for _ in range(args.get("max_pages", 10)):
req = urllib.request.Request(
url, headers={"User-Agent": USER_AGENT, "Accept": "application/json"})
with urllib.request.urlopen(req, timeout=30, context=_ssl_context()) as resp:
data = json.loads(resp.read().decode("utf-8", "replace"))
jobs = []
for it in data.get("items", []):
items = data.get("items", []) or []
for it in items:
# Prefer the schema.org JobPosting payload's real address over default_location.
# Axpo is a pan-European trader (Madrid, Milan, Warsaw, Germany, France as well as
# Baden/Zurich), so defaulting every role to the board's home country would mark
# ~380 non-Swiss roles CH-eligible — the same class of bug as Apple's forced
# "Switzerland" default. Telenor's feed has no _jobposting and falls back cleanly.
location = args.get("default_location", "")
places = ((it.get("_jobposting") or {}).get("jobLocation") or [])
if isinstance(places, dict):
places = [places]
parts = []
for place in places:
addr = (place or {}).get("address") or {}
# schema.org gives ISO alpha-2 ("CH", "NO"); the location policy matches on
# full country names, so expand or the role reads as location-unknown.
country = _ISO2_COUNTRY.get((addr.get("addressCountry") or "").upper(),
addr.get("addressCountry"))
bit = ", ".join(x for x in (addr.get("addressLocality"), country) if x)
if bit and bit not in parts:
parts.append(bit)
if parts:
location = " | ".join(parts)
jobs.append({
"id": it.get("id") or it.get("url", ""),
"title": it.get("title", ""),
"location": args.get("default_location", ""),
"location": location,
"url": it.get("url", ""),
"posted": it.get("date_published", ""),
"description": re.sub(r"<[^>]+>", " ", it.get("content_html", ""))[:2500],
})
nxt = data.get("next_url")
if not items or not nxt or nxt == url:
break
url = nxt
titles = {j["title"].strip().lower() for j in jobs}
if jobs and len(titles & DEMO_TITLES) >= 4:
raise RuntimeError(
@@ -685,9 +857,10 @@ def fetch_teamtailor(args):
def fetch_getro(args):
"""Getro network job-board search API (POST JSON). Powers VC portfolio talent
networks — here the Coinbase Ventures web3 network (collection 1625). Returns roles
across ALL portfolio companies (Notion, Ashby, VALR, World, ...), NOT Coinbase itself;
Coinbase doesn't list its own openings on its Ventures board. Server-side filters:
networks. NO company in COMPANIES currently uses this adapter — the Coinbase Ventures
network was dropped 2026-08-18 (see the note there). Kept because it generalises to any
Getro collection id. Note such boards list PORTFOLIO companies' roles, never the VC's
own openings, so an org filter is usually needed. Server-side filters:
searchable_locations and job_functions. Org name is folded into the title since this
is a multi-company board."""
collection = args["collection"]
@@ -731,6 +904,94 @@ def fetch_getro(args):
return jobs
def fetch_amazon(args):
"""amazon.jobs public search JSON. Covers Amazon + AWS, which share one board.
`normalized_country_code[]` is the filter that actually works; the plain `country[]`
and `loc_query` params are ignored and silently return the global (US-heavy) set —
verified 2026-08-18, where `country[]=CHE` returned 6,837 mostly-Seattle hits against
32 for the normalized form."""
base = "https://www.amazon.jobs/en/search.json"
headers = {"User-Agent": ("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
"(KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36")}
jobs, offset, page_size = [], 0, 100
while True:
qs = urllib.parse.urlencode({
"radius": "100mi", "offset": offset, "result_limit": page_size,
"sort": "recent", "base_query": args.get("base_query", ""),
})
countries = "".join(f"&normalized_country_code[]={urllib.parse.quote(c)}"
for c in args.get("countries", ["CHE"]))
data = http_get_json(f"{base}?{qs}{countries}", headers=headers)
batch = data.get("jobs", []) or []
for j in batch:
jobs.append({
"id": str(j.get("id_icims") or j.get("id") or ""),
"title": j.get("title", ""),
"location": j.get("normalized_location") or j.get("location") or "",
"url": "https://www.amazon.jobs" + (j.get("job_path") or ""),
"posted": j.get("posted_date", ""),
"description": (j.get("description_short") or j.get("description") or "")[:2000],
})
total = data.get("hits", 0)
offset += page_size
if not batch or offset >= total or offset >= args.get("max_results", 500):
break
return jobs
def fetch_phenom(args):
"""Phenom People careers search (POST /widgets with ddoKey=refineSearch). Used by Roche.
The public search-results page is a JS shell whose location filter silently fails: a
scrape of `?locationsearch=Switzerland` returns *recommendation-widget* cards (Shanghai,
Kyiv, Bogota) while the page state reports `no-results`. Verified 2026-08-18 — this is
why Roche contributed 0 Swiss roles for months while looking healthy. The underlying
refineSearch endpoint filters correctly (134 CH roles), so query it directly.
`country_facet` is the value for the `country` facet (e.g. "Switzerland"). `ref_num` and
`page_id` are tenant constants visible in any /widgets POST from the careers site."""
url = args["url"]
ref_num, page_id = args["ref_num"], args.get("page_id", "page11-ds")
lang = args.get("lang", "en_global")
page_size = args.get("page_size", 100)
headers = {
"Origin": "{0.scheme}://{0.netloc}".format(urllib.parse.urlsplit(url)),
"Referer": url,
# Phenom 403s the default library UA.
"User-Agent": ("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
"(KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"),
}
jobs, offset = [], 0
while True:
data = http_get_json(url, method="POST", headers=headers, data={
"lang": lang, "deviceType": "desktop", "country": "global",
"pageName": "search-results", "ddoKey": "refineSearch",
"sortBy": "", "subsearch": "", "from": offset, "jobs": True, "counts": True,
"all_fields": ["country", "state", "city", "category", "type"],
"size": page_size, "clearAll": False, "jdsource": "facets",
"isSliderEnable": False, "pageId": page_id, "siteType": "external",
"keywords": "", "global": True, "locationData": {},
"selected_fields": {"country": [args["country_facet"]]},
})
res = data.get("refineSearch", {}) or {}
batch = (res.get("data", {}) or {}).get("jobs", []) or []
for j in batch:
jobs.append({
"id": str(j.get("jobId") or j.get("jobSeqNo") or ""),
"title": j.get("title", ""),
"location": j.get("location") or j.get("cityStateCountry") or "",
"url": j.get("applyUrl", ""),
"posted": (j.get("postedDate") or "")[:10],
"description": (j.get("descriptionTeaser") or "")[:2000],
})
total = res.get("totalHits", 0)
offset += page_size
if not batch or offset >= total or offset >= args.get("max_results", 600):
break
return jobs
def fetch_onlyfy(args):
"""onlyfy.jobs board (XING E-Recruiting / ex-Prinzip), used by Bitcoin Suisse. The
candidate/job/ajax_list endpoint returns an HTML fragment listing every posting; each
@@ -1315,6 +1576,8 @@ ADAPTERS = {
"rss": fetch_rss,
"teamtailor": fetch_teamtailor,
"getro": fetch_getro,
"amazon": fetch_amazon,
"phenom": fetch_phenom,
"onlyfy": fetch_onlyfy,
"lever": fetch_lever,
"taleo": fetch_taleo,
@@ -1429,15 +1692,111 @@ def save_seen(seen):
_atomic_write_json(STATE_FILE, seen)
# Board-specific req-ID patterns, tried in order against the URL path+query. Every key is
# scoped by host (see _decision_key), so an ID only ever has to be unique within one board
# and cross-board collisions are impossible by construction.
_JOB_ID_PATTERNS = [
# Amazon: the slug is truncated to a fixed width and that width has changed between runs,
# so the same req yields different URLs. This is the drift that motivated the whole helper.
(r"amazon\.jobs$", r"/jobs/(\d+)"),
(r"careers\.microsoft\.com$", r"/job/(\d+)"),
(r"google\.com$", r"/results/(\d+)-"),
# Workday: trailing _REQID on the job slug (Roche 202608-120791-1, Novartis REQ-…, NVIDIA JR…).
# Also collapses the /apply suffix some rows carry and others don't.
(r"myworkdayjobs\.com$", r"_([A-Za-z]*[\d-]+\d)(?:/apply)?/?$"),
(r"smartrecruiters\.com$", r"/(\d{6,})"),
(r"greenhouse\.io$", r"/jobs/(\d+)"),
(r"careers\.cisco\.com$", r"/job/(\d+)"),
# PostFinance publishes one req per locale (…/74421-fr_FR, …/74372-de_DE) — same job.
(r"jobs\.postfinance\.ch$", r"/(\d+)-[a-z]{2}_[A-Z]{2}/?$"),
# Apple: same req appears with ?team=… and with a /locationPicker suffix.
(r"jobs\.apple\.com$", r"/details/([\d-]+)"),
(r"successfactors\.eu$", r"[?&]jobId=(\d+)"),
(r"taleo\.net$", r"[?&]job=(\d+)"),
(r"bis\.org$", r"/vacancies/(jr\d+)"),
(r"metacareers\.com$", r"/job_details/(\d+)"),
(r"finn\.no$", r"[?&]finnkode=(\d+)"),
(r"linkedin\.com$", r"/jobs/view/.*-(\d+)/?$"),
(r"careers\.roche\.com$", r"/job/([\w-]+)/"),
# Greenhouse-backed careers pages that carry the req in a query param (Databricks,
# Fivetran, Datadog, Elastic — Elastic repeats gh_jid twice; the first match wins).
(None, r"[?&]gh_jid=(\d+)"),
# Teamtailor (Axpo, Telenor) — /jobs/<id>-<slug>.
(None, r"/jobs/(\d+)-"),
# Ashby / Lever / Recruitee-style boards (RUAG, BKW, BFH, SBB) — trailing UUID.
(None, r"/([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-"
r"[0-9a-f]{4}-[0-9a-f]{12})/?$"),
]
def _decision_key(url):
"""Stable per-job identity for the decision log, resilient to URL drift.
Boards rewrite their own URLs: Amazon truncates the slug to a width that has changed
between runs, PostFinance publishes one URL per locale, Apple appends ?team= or
/locationPicker, Workday adds /apply. Keying decisions on the raw URL therefore loses
them — an already-skipped role silently resurfaces as undecided (this is what happened
to the AWS Principal Delivery Consultant req between the 2026-08-18 and 08-21 runs).
Returns "<host>#<req-id>" when a board pattern matches, else "<host>#<normalized path>".
Host-scoping keeps two boards that use the same numbering scheme from colliding, so a
wrong match can never mark a genuinely undecided role as decided at a *different* company.
"""
try:
parts = urllib.parse.urlsplit(url.strip())
except ValueError:
return url.strip()
host = (parts.netloc or "").lower().split(":")[0]
if host.startswith("www."):
host = host[4:]
target = parts.path + (f"?{parts.query}" if parts.query else "")
for host_pat, id_pat in _JOB_ID_PATTERNS:
if host_pat and not re.search(host_pat, host):
continue
m = re.search(id_pat, target)
if m:
return f"{host}#{m.group(1)}"
# No known ID shape: fall back to the path with cosmetic variation stripped.
path = parts.path.rstrip("/")
if path.endswith("/apply"):
path = path[: -len("/apply")]
return f"{host}#{path.lower()}"
def load_decisions():
"""Decision log keyed by job URL: {url: {company, title, decision, note, date}}.
Decisions persist across runs so we don't re-evaluate roles we've already judged
(shortlist / skip / applied / paused / rejected — free-text, not enforced)."""
(shortlist / skip / applied / paused / rejected — free-text, not enforced).
The file stays URL-keyed — it is hand-edited and the URL is the useful thing to read.
Lookups go through index_decisions() instead, which is drift-proof."""
if DECISIONS_FILE.exists():
return json.loads(DECISIONS_FILE.read_text(encoding="utf-8"))
return {}
def index_decisions(decisions):
"""Build {stable key -> entry} for lookup. On collision the newest decision wins."""
idx = {}
for url, entry in decisions.items():
key = _decision_key(url)
prev = idx.get(key)
if prev is None or str(entry.get("date", "")) >= str(prev.get("date", "")):
idx[key] = entry
return idx
def find_decision_url(decisions, url):
"""Existing log URL for the same job, if one is already recorded under a drifted URL."""
key = _decision_key(url)
for existing in decisions:
if _decision_key(existing) == key:
return existing
return None
def save_decisions(decisions):
_atomic_write_json(DECISIONS_FILE, decisions)
@@ -1507,6 +1866,8 @@ def write_stats_table(stats, total_secs):
def write_report(path, results, errors, new_only, include_weak, stats=None, total_secs=0.0,
decisions=None, hide_decided=False):
# `decisions` here is the stable-key index from index_decisions(), not the raw
# URL-keyed log — job URLs drift between runs, the keys do not.
decisions = decisions or {}
today = datetime.now().strftime("%Y-%m-%d")
n_new = sum(1 for r in results if r["is_new"])
@@ -1531,7 +1892,7 @@ def write_report(path, results, errors, new_only, include_weak, stats=None, tota
if not include_weak and weak:
lines.append(f"\n_Hiding {len(weak)} weak/noise roles (score < 2). Use --include-weak to show._")
n_decided = sum(1 for r in results if r["url"] in decisions)
n_decided = sum(1 for r in results if _decision_key(r["url"]) in decisions)
if n_decided:
shown = "hidden" if hide_decided else "tagged inline"
lines.append(f"_{n_decided} role(s) already in the decision log ({shown}; "
@@ -1543,12 +1904,13 @@ def write_report(path, results, errors, new_only, include_weak, stats=None, tota
buckets.append(("Weak / noise (score < 2)", weak))
for bucket_name, bucket in buckets:
shown = [r for r in bucket if not (hide_decided and r["url"] in decisions)]
shown = [r for r in bucket
if not (hide_decided and _decision_key(r["url"]) in decisions)]
if not shown:
continue
lines.append(f"\n## {bucket_name} - {len(shown)} role(s)\n")
for r in shown:
d = decisions.get(r["url"])
d = decisions.get(_decision_key(r["url"]))
new_tag = " [NEW]" if r["is_new"] else ""
decided_tag = f" — 🗂 {d['decision'].upper()}" if d else ""
loc_tag = ("CH" if r["in_ch"] else
@@ -1613,10 +1975,16 @@ def _process_company(cid, display, args, jobs, seen, today, last_scrape):
"0 jobs returned (verify board slug/selectors if this is unexpected)",
))
# Inclusion allowlist (fails closed — only for boards too large to score in full).
title_filter = args.get("_title_filter")
if title_filter:
jobs = [j for j in jobs
if any(_kw_in(k, (j.get("title") or "").lower()) for k in title_filter)]
# Exclusion denylist (fails open — preferred; see NOISE_TITLE_EXCLUDE).
title_exclude = args.get("_title_exclude")
if title_exclude:
jobs = [j for j in jobs
if not any(_kw_in(k, (j.get("title") or "").lower()) for k in title_exclude)]
dates = [d for j in jobs if (d := _parse_posted(j.get("posted")))]
newest = max(dates) if dates else None
@@ -1678,7 +2046,10 @@ def main():
return
url, status, note = rest[0], rest[1], " ".join(rest[2:])
decisions = load_decisions()
prev = decisions.get(url, {})
# Update in place if this job is already logged under a drifted URL (truncated
# slug, other locale, /apply suffix) rather than adding a second row for it.
existing = find_decision_url(decisions, url)
prev = decisions.pop(existing, {}) if existing else {}
# Preserve any company/title already stored (filled when a report run tagged the URL).
decisions[url] = {
"company": prev.get("company", ""),
@@ -1686,6 +2057,8 @@ def main():
"decision": status, "note": note, "date": today,
}
save_decisions(decisions)
if existing and existing != url:
print(f"(replaced earlier URL for the same job: {existing})", file=sys.stderr)
print(f"Recorded: {status} — {url}", file=sys.stderr)
return
@@ -1712,6 +2085,7 @@ def main():
seen = load_seen()
decisions = load_decisions()
dec_index = index_decisions(decisions)
last_scrape = load_last_scrape()
all_results, errors, stats = [], [], []
new_last_scrape = dict(last_scrape)
@@ -1791,7 +2165,7 @@ def main():
report_path = REPORTS_DIR / f"{today}.md"
write_report(report_path, all_results, errors, new_only, include_weak,
stats=stats, total_secs=total_secs,
decisions=decisions, hide_decided=hide_decided)
decisions=dec_index, hide_decided=hide_decided)
# Plain data dump of every eligible role, unfiltered by keyword score — fit judgment
# for these is done in conversation against the profile, not by the keyword scorer.
@@ -1799,7 +2173,7 @@ def main():
dump = [{
"company": r["company"], "title": r["title"], "location": r["location"],
"url": r["url"], "posted": r["posted"], "is_new": r["is_new"],
"decision": decisions.get(r["url"]),
"decision": dec_index.get(_decision_key(r["url"])),
"keyword_signal": {"score": r["score"], "pos": r["pos"], "neg": r["neg"]},
"loc_flags": {
"in_ch": r["in_ch"], "remote": r["remote"], "relocation": r["relocation"],
@@ -1826,18 +2200,21 @@ def main():
# Automated (COMPANIES above):
# workday nvidia, novartis, equinor (norway_only)
# ashby kraken, openai, confluent, snowflake
# greenhouse anthropic, gitlab, grafana, databricks, datadog, elastic, dbtlabs
# greenhouse anthropic, gitlab, grafana, databricks, datadog, elastic, fivetran
# pcsx microsoft
# smartrecruiters metgroup, ldc
# smartrecruiters metgroup, cern
# rss bis
# getro coinbase_ventures
# amazon amazon (Amazon + AWS share one board)
# phenom roche (Cisco still uses the playwright path)
# teamtailor telenor, axpo (custom domain via base_url)
# getro (no company configured — adapter kept for future VC talent networks)
# onlyfy bitcoin_suisse
# lever palantir, quantco
# taleo nato (dk_no; nato.taleo.net §2 — marketing page is nato.int/vacancies)
# json swissgrid
# sbb sbb
# bkw bkw
# playwright google, apple, meta, roche, cisco, ruag, postfinance, bfh
# playwright google, apple, meta, cisco, ruag, postfinance, bfh
#
# MANUAL_CHECK: Oracle (ORC needs CH geographyId).
# ==============================================================================
job_scout/state/application_cohort.json
+90
View File
@@ -0,0 +1,90 @@
{
"cohort_id": "evidence-first-2026-01",
"status": "active",
"started": "2026-07-27",
"target_mix": {
"core": 7,
"adjacent": 2,
"stretch": 1
},
"definitions": {
"core": "Evidence Fit 75+ and no hard-gate failure",
"adjacent": "Evidence Fit 60-74, no required Gap, normally at most one serious preferred gap",
"stretch": "Evidence Fit below 60 or a defining responsibility is only Adjacent"
},
"applications": [
{
"company": "Google",
"role": "Software Engineer III, Business Home",
"application_date": "2026-07-27",
"fit_class": "core",
"evidence_fit": 89.0,
"hard_gate": "pass",
"channel": "weak",
"outcome": "rejected_no_interview"
},
{
"company": "Aker BP ASA",
"role": "Data Product Architect - AI-ready Data Products",
"application_date": "2026-07-29",
"fit_class": "adjacent",
"evidence_fit": 79.0,
"hard_gate": "pass",
"channel": "weak",
"outcome": "rejected_no_interview",
"submitted": "2026-07-29",
"rejection_date": "2026-08-27"
},
{
"company": "SBB",
"role": "Data Engineer (m/w/d), Asset Management Infrastrukturanlagen, Bern",
"job_id": "103755",
"application_date": "2026-08-25",
"fit_class": "core",
"evidence_fit": 79.0,
"hard_gate": "pass",
"channel": "weak",
"outcome": "interview_invited",
"submitted": "2026-08-25",
"open_risk": "Anforderungsniveau K is a capped GAV band, figures not public; seat may read lateral or below current Staff + Component Owner scope.",
"interview": {
"stage": "Stage 2 of 4 - Virtuelles Kennenlernen mit HR und Fuehrungskraft",
"invited": "2026-08-28",
"scheduled": "2026-09-09 08:30",
"duration_min": 45,
"format": "Microsoft Teams"
},
"note": "FIRST interview invitation in the evidence-first cohort. Converted from a COLD submit - the published warm contact (Andri Wienandts, +41 79 364 62 53) was never called. The resume never contained Snowflake, dbt or Power BI, the three literals an ATS keyword screen would have keyed on, so a human read the dossier."
},
{
"company": "Citadel Securities",
"role": "Platform Engineer, Research Platform Engineering (Zurich/NY/Miami)",
"application_date": "2026-08-26",
"fit_class": "adjacent",
"evidence_fit": 71.0,
"hard_gate": "pass",
"channel": "weak",
"outcome": "applied"
},
{
"company": "Schweizer Armee - Kommando Cyber",
"role": "DevOps Engineer III (Data Platform)",
"application_date": "2026-08-27",
"fit_class": "core",
"evidence_fit": 79.0,
"hard_gate": "pass",
"channel": "weak",
"outcome": "applied"
},
{
"company": "RUAG MRO Holding AG (Business Area C5I)",
"role": "AI Engineer C5I (w/m/d), Thun",
"application_date": "2026-08-27",
"fit_class": "stretch",
"evidence_fit": 74.0,
"hard_gate": "fail",
"channel": "weak",
"outcome": "applied"
}
]
}
job_scout/state/decisions.json
+99 -22
View File
@@ -125,13 +125,6 @@
"note": "C++/automotive, weaker fit vs his Python/data-platform stack.",
"date": "2026-06-01"
},
"https://jobs.ashbyhq.com/snowflake/3eea87fa-73c8-46dc-b69b-7beb438b48d8": {
"company": "Snowflake",
"title": "Senior Software Engineer, Enterprise (Observe by Snowflake)",
"decision": "applied",
"note": "Observability-spine SWE on an Iceberg telemetry lakehouse; Zurich-commutable, comp CHF 176-253k base clears bar. Best-fit role in the 2026-06 search. | SENT 2026-06-06 (resume+CL finalized, ~86/100).",
"date": "2026-06-08"
},
"https://www.google.com/about/careers/applications/jobs/results/87066954308690630-senior-data-engineer?location=Switzerland": {
"company": "Google",
"title": "Senior Data Engineer (Merchant Data Science)",
@@ -163,9 +156,9 @@
"https://job-boards.greenhouse.io/gitlab/jobs/8522408002": {
"company": "GitLab",
"title": "Forward Deployed Engineer - Germany (Staff)",
"decision": "shortlist",
"note": "Remote-CH eligible Staff FDE (Duo Agent Platform adoption, regulated/self-managed envs). CAUTION: JD body says 'strategic accounts in the APJ region' despite Germany title — clarify region/timezone with recruiter before tailoring. Gaps: Ruby/Go ideal, Terraform/Ansible/Helm expected (Kraken-style honest gap), GitLab-internals depth. Travel up to 50%.",
"date": "2026-07-03"
"decision": "skip",
"note": "HARD GATE FAIL: 'Strong Ruby on Rails or Go experience and the ability to contribute directly to GitLab product code.' Stack is Python/Java. JD explicitly states this is NOT a generic field-engineering role - it is a product-codebase contributor role. Terraform also listed (known honest gap).",
"date": "2026-08-18"
},
"https://www.google.com/about/careers/applications/jobs/results/135155660865053382-staff-research-generative-ai-cloud-ai-research-coscientist?location=Switzerland": {
"company": "Google",
@@ -1574,13 +1567,6 @@
"note": "Claude judgment 2026-07-06: same Observe team as your pending application, but this req (query-execution internals) is more database-internals-specialized than your general profile",
"date": "2026-07-06"
},
"https://jobs.ashbyhq.com/snowflake/db4f0492-ae21-49b7-a7eb-100be61e92bb": {
"company": "Snowflake",
"title": "Senior Software Engineer - Observe by Snowflake, Metrics Platform",
"decision": "paused",
"note": "Strong Observe metrics/data-platform overlap and CHF 176-253k band, but an Enterprise-team application has been open since 2026-06-06. Do not stack another cold application to the same organization/team; pursue recruiter routing or wait for status.",
"date": "2026-07-28"
},
"https://jobs.ashbyhq.com/snowflake/26a0ae52-97a6-4a46-9216-3c382570d89b": {
"company": "Snowflake",
"title": "Software Engineer Intern - Zurich (2026)",
@@ -1715,11 +1701,11 @@
"date": "2026-07-28"
},
"https://www.finn.no/job/fulltime/ad.html?finnkode=469067315": {
"company": "",
"title": "",
"decision": "applied",
"note": "Aker BP ASA - Data Product Architect, AI-ready Data Products (FINN 469067315). SUBMITTED 2026-07-30, ahead of 2026-08-02 deadline. Evidence Fit 79/100, Document Quality 93/100, hard gate PASS, channel weak. Stavanger/Oslo/Trondheim, English working language.",
"date": "2026-07-31"
"company": "Aker BP ASA",
"title": "Data Product Architect, AI-ready Data Products",
"decision": "rejected",
"note": "Aker BP ASA - Data Product Architect, AI-ready Data Products (FINN 469067315). SUBMITTED 2026-07-30, ahead of 2026-08-02 deadline. Evidence Fit 79/100, Document Quality 93/100, hard gate PASS, channel weak. Stavanger/Oslo/Trondheim, English working language. REJECTED 2026-08-27, no interview - ~28 days to disposition (records differ on the submission date: session file and cohort say 2026-07-29, this note and CLAUDE.md say 2026-07-30). Cold channel; the optional follow-up email to Per Olav Marthinsen stayed an open action with no record of it being sent. Cohort evidence-first-2026-01 slot stays consumed (adjacent 1 of 2); the rejection does not free it. Norway remains an active, selective lane - NATO JWC Stavanger is still open, and the user has said Equinor and other strong Norwegian employers stay in scope.",
"date": "2026-08-27"
},
"https://www.google.com/about/careers/applications/jobs/results/93922217108087494-software-engineer-iii-business-home?location=Switzerland": {
"company": "Google",
@@ -1767,5 +1753,96 @@
"decision": "skip",
"note": "CLOSED - req taken down. Live Playwright scrape 2026-08-11 returns 'Job not found. This job may have been taken down.'; also absent from the current Google Zurich board listing (32 rows). Never applied.",
"date": "2026-08-11"
},
"https://www.amazon.jobs/en/jobs/10504263/senior-forward-deployed-engineer-aws-forward-deployed-engineering": {
"company": "",
"title": "",
"decision": "skip",
"note": "Phase 0 done, then NO-GO 2026-08-18. Evidence Fit 69/100. Minimum quals all Direct, but customer embedding is title-defining and globally forbidden in claims.json; multi-agent/retrieval is a second Gap. User bar: not worth a package below ~75. Do not re-surface this req.",
"date": "2026-08-18"
},
"https://www.amazon.jobs/en/jobs/10450340/senior-proserve-cloud-architect-healthcare-and-life-sciences-hcls-awsi-sales": {
"company": "",
"title": "",
"decision": "skip",
"note": "Bern location is attractive but the domain gate is clinical pharma: CDISC SDTM/ADaM, eCTD, GxP, CTMS, EDC, regulatory submissions. No evidence in any of that. Location alone does not carry it.",
"date": "2026-08-18"
},
"https://www.amazon.jobs/en/jobs/10478293/principal-delivery-consultant-technical-lead-genai-ml-data-science-professional-servic": {
"company": "",
"title": "",
"decision": "skip",
"note": "Scored highest (7) but that is the SA/architect overscoring trap. Basic quals require 8+ yrs enterprise architecture delivery INCLUDING 5+ yrs leading technical teams governing architecture on transformation programmes, plus VP+ executive communication. Level, people-leadership and exec-facing gaps compound; preferred quals add HCLS pharma + vector DB/RAG.",
"date": "2026-08-18"
},
"https://jobs.careers.microsoft.com/global/en/job/200044133": {
"company": "",
"title": "",
"decision": "skip",
"note": "MISLABELLED REQ. Title says Principal Forward Deployed Engineer - Data Scientist - German Speaking, but required quals are network engineering: BGP, MPLS, SD-WAN, IPv4/IPv6, Palo Alto firewalls, Azure Virtual WAN, DNS/DHCP, load balancers, IDS/IPS, Zero Trust. This is a network security/infra role. Complete mismatch - do not trust the title on Microsoft FDE siblings.",
"date": "2026-08-18"
},
"https://jobs.sbb.ch/v2/offene-stellen/data-engineer-m-w-d-asset-management-infrastrukturanlagen/45fa2411-31dc-4520-827e-1c93a527534c": {
"company": "",
"title": "",
"decision": "interview",
"note": "INTERVIEW INVITED 2026-08-28 for 2026-09-09 08:30, 45 min Teams. Stage 2 of 4 per the posting: \"Virtuelles Kennenlernen mit HR und Fuehrungskraft\" - so HR and Andri Wienandts together, not a pure HR screen. Submitted 2026-08-25, Core, Evidence Fit 79/100, Document Quality 92/100, cold channel. First interview of the evidence-first cohort. Unresolved and now live: Anforderungsniveau K band, design/architecture scope vs current Staff + Component Owner level, RAMSI Java/Spring Boot/Angular share of the role, Kidz Care actual rate.",
"date": "2026-08-28"
},
"https://jobs.ashbyhq.com/snowflake/3eea87fa-73c8-46dc-b69b-7beb438b48d8": {
"company": "Snowflake",
"title": "Senior Software Engineer, Enterprise (Observe by Snowflake)",
"decision": "closed_no_response",
"note": "CLOSED 2026-08-25 by user decision at 80 days silent (sent 2026-06-06, ~86/100, 2pp resume + 1pp CL, cold channel). NO rejection was ever received - this is a presumed-dead call, not a recorded rejection; do not count it in rejection statistics. 80 days exceeds the longest real disposition in the log (Equinor, 79). Pre-dates the evidence-first-2026-01 cohort, so no cohort slot is freed.",
"date": "2026-08-25"
},
"https://jobs.ashbyhq.com/snowflake/db4f0492-ae21-49b7-a7eb-100be61e92bb": {
"company": "Snowflake",
"title": "Senior Software Engineer - Observe by Snowflake, Metrics Platform",
"decision": "maybe",
"note": "UNPAUSED 2026-08-25: the blocking reason is gone - the Enterprise/Observe application it was stacked behind was closed no-response at 80 days. Strong Observe metrics/data-platform overlap, CHF 176-253k band, Zurich-commutable. Caveat: the prior cold submit to this org drew zero response despite an ~86/100 package, so do not simply cold-apply again - route via recruiter or a warm contact.",
"date": "2026-08-25"
},
"https://www.citadelsecurities.com/careers/details/platform-engineer/": {
"company": "",
"title": "",
"decision": "applied",
"note": "Submitted 2026-08-26. Adjacent, Evidence Fit 71, Document Quality ~91 est (critique scored 86 pre-Tier-1). 2pp resume + 1pp CL, both validator PASS. IMPORTANT FINDING: the application form's preferred-work-location dropdown offered NO Zurich option, despite the JD body saying 'Miami, Zurich or New York' and the site's own Zurich filter returning this req. Careers site is operated by Citadel Enterprise Americas LLC. Treat the Zurich seat as unconfirmed for any future Citadel application - check the form before investing. Working-model question (5-day onsite?) never resolved and may now be moot. Channel cold; user applied with low expectation.",
"date": "2026-08-26"
},
"https://careers.axpo.com/jobs/8263436-staff-plattform-engineer-w-m-d": {
"company": "",
"title": "",
"decision": "skip",
"note": "User decision 2026-08-27: not of interest. Baden is a 90min commute from Thun and Axpo is unlikely to pay materially above his current ~140k. RULE: for a commute that long the package must be much better - BIS-style tax-free or FAANG-like comp. Azure + Terraform are also named requirements with 0 canonical evidence, but that was NOT the deciding factor.",
"date": "2026-08-27"
},
"https://jobs.ruag.ch/offene-stellen/ai-engineer-c5i/4afaa9aa-74c1-4b4d-9232-cf65651ce8ec": {
"company": "RUAG C5I",
"title": "AI Engineer C5I (Thun)",
"decision": "applied",
"note": "SUBMITTED 2026-08-27 via jobs.ruag.ch (portal only; email and post are refused). Application ID 18027. Cohort evidence-first-2026-01 now 6/10 and the sole Stretch slot is CONSUMED (stretch 1/1). Evidence Fit 74/100, strategic Stretch, strict title-capability hard gate FAIL - all knowingly accepted under the user's explicit Phase 0 override; R1 (build/operate an internal AI-/LLM platform incl. compute cluster) and R2 (in-house LLM solutions, connectors, retrieval/document solutions) are Adjacent, not Direct. Document Quality 93/100 after a full critique-and-fix round (85 pre-edit), truth and provenance 24/25, zero Tier 1 findings at Round 2. Submitted package: German Swiss/DACH resume, 2 pages, 13 bullets, plus a German motivation letter, 1 page, 295 words. Both validators PASS with 0 warnings, 0 boxes, clean two-pass compiles, text-order checks and visual QA of every page. Positioning: Staff Data, Analytics & AI Engineer (his canonical title) with production ML-inference integration, Kubernetes/Linux/Ansible automation, current LLM application configuration and the canonical BW-1 officer career. PP-3 (private self-hosted Debian server, hardening) carries the current Linux/ops evidence since BS-6 ended Dec 2022. The letter names the AI-platform gap in one sentence and pivots to the operations/platform side, and connects SW-5 (Security Champion 2025/2026) to C5I's stated DevSecOps model. No RAG, retrieval, compute-cluster, GPU, model-training or LLM-platform ownership claim anywhere. Channel WEAK - cold portal submit; Marco Heinzen's published direct line (+41 79 568 14 96) was never used. STILL UNRESOLVED and now live screening topics: compensation (no band published; external samples suggest downside risk against the 180k bar), PSP/project eligibility for a German citizen with a Swiss B permit, and whether the role is lateral or a down-level move against current Staff + Component Owner scope.",
"date": "2026-08-27"
},
"https://nvidia.wd5.myworkdayjobs.com/job/Switzerland-Zurich/Senior-Site-Reliability-Engineer--DGX-Cloud_JR2021427-1": {
"company": "",
"title": "",
"decision": "shortlist",
"note": "User flagged interesting 2026-08-27. NVIDIA JR2021427, posted 2026-08-20. KEY: Workday additionalLocations lists 'Switzerland, Remote' alongside Zurich - the commute constraint that killed Axpo does not apply, which re-ranks this level with RUAG on practical grounds. Comp is the FAANG-like shape he named as worth a trip anyway. Fit: K8s at scale, Python, observability (Prometheus/Grafana/ELK) Direct; GPU cluster ops across AWS/GCP/Azure/OCI. GAPS, both named requirements with 0 canonical evidence: (a) infrastructure automation - Terraform/Ansible/Chef/Puppet, Terraform is a recorded hard gap and the Kraken SRE rejection was exactly this shape; (b) SRE practice as a discipline - SLO/SLI/error budgets, on-call rotation, incident command, no SRE title. Also '10+ years operating production services'. NOTE: the scout's constructed URL 404s - the real apply link needs the NVIDIAExternalCareerSite segment: https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite/job/Switzerland-Zurich/Senior-Site-Reliability-Engineer--DGX-Cloud_JR2021427-1",
"date": "2026-08-27"
},
"https://www.bis.org/careers/vacancies/jr100420.htm": {
"company": "",
"title": "",
"decision": "maybe",
"note": "User flagged interesting 2026-08-27; NOT recorded as shortlist because the JD read is worse than the title suggested. BIS Basel, Platforms team in ITS, 3-year fixed term, deadline 17 Sept 2026, hires all nationalities with relocation support, English corporate language - all fine and consistent with jr100429. PROBLEM: the title-defining capability is DBA, not platform data engineering. Purpose line is enterprise database platforms - Microsoft SQL Server, PostgreSQL, Oracle - with 'deep domain expertise in database technologies', DB administration/availability/lifecycle, and DB certifications as an advantage. That is a Gap, not a stretch. Same shape as the AWS Senior FDE NO-GO on 2026-08-18: minimum quals largely pass (K8s, IaC, CI/CD, observability, cloud all Direct) but the capability the title is built on is unsupported. Secondary issue: jr100429 has been open at BIS since 2026-07-10, so this would stack a second application at one org - the Snowflake precedent. Deadline is 3 weeks out, no urgency pressure.",
"date": "2026-08-27"
},
"https://jobs.admin.ch/offene-stellen/devops-engineer-iii-data-platform/752ded37-1753-4ac8-946a-b0c958d4424a": {
"company": "Schweizer Armee - Kommando Cyber",
"title": "DevOps Engineer III (Data Platform)",
"decision": "applied",
"note": "Submitted 2026-08-27. Ref JRQ$540-19848, Zimmerwald. Core, Evidence Fit 79, Document Quality 92, hard gate PASS, channel weak. German 2-page CV plus 1-page motivation letter; the six-year Bundeswehr officer career is included as military-context and motivation evidence. User explicitly accepted the critique's remaining BS-1 wording issue ('ohne manuellen Eingriff') as good enough and submitted without the proposed correction. Honest gaps at submission: physical datacenter work, Pull-GitOps as a named practice, GPGPU, ClickHouse and second Swiss official language. Lohnklasse and Engineer-III level were not recorded as clarified.",
"date": "2026-08-27"
}
}
knowledge_base/extractions/thiessen_certifications.md
+2 -1
View File
@@ -1,7 +1,7 @@
# Dennis Thiessen — Certifications Extraction
> One file for all cert PDFs. Added as each cert is processed.
> Last updated: 2026-03-28
> Last updated: 2026-08-27
---
@@ -16,6 +16,7 @@
| 5 | Projektleiter Baustein A/IT — Das IT-Projektmanagement (seminar) | Integrata AG | 1–5 Sep 2014 | N/A | Seminar-Nr. 2113 | 5-day attendance (Teilnahmebestätigung). NOT a cert — do not list on resume/CV. Topics: IT project planning, control, risk analysis. Context: during Bundeswehr period. |
| 6 | AI for Trading Nanodegree | Udacity (co-created with WorldQuant) | 13 May 2021 | N/A | confirm.udacity.com/DJ9QTAH9 | Quantitative finance + ML for trading; completed during Bosch period. Relevant for quant/fintech roles; niche signal for blockchain/crypto interest. |
| 7 | AWS Certified Solutions Architect – Associate | Amazon Web Services (via Alpine Testing / CertMetrics) | 26 Sep 2024 | 26 Sep 2027 (active) | cp.certmetrics.com/amazon — verified 2026-03-27 | ACTIVE. High-value cert for Data Platform / Infra and Staff Data Engineer roles. Pairs well with Udacity DataEng AWS nanodegree. |
| 8 | IBM AI Engineering (Professional Certificate, 6 courses) | IBM via Coursera | 4 Jun 2020 | N/A (no expiry) | coursera.org/verify/professional-cert/3ZBZFVAL6A34 | Name: "Dennis Thießen" (ß variant). **Verified against the certificate PDF 2026-08-27** (`C:\myCloud\Bewerbungsunterlagen\Zeugnisse\Zertifikate\cert_IBM_AI_Engineering.pdf`). Courses: Machine Learning with Python; Scalable ML on Big Data with Apache Spark; Intro to Deep Learning & Neural Networks with Keras; Deep Neural Networks with **PyTorch**; Building Deep Learning Models with **TensorFlow**; AI Capstone Project with Deep Learning. This is the source of the `TensorFlow/Keras` **and** `PyTorch` entries in `claims.json` — both stay `certification-context-only`. Explicitly non-credit: the certificate states it confers no grade, course credit or degree, so never present it as an academic qualification. |
---
output/AWS_Senior_FDE_Zurich/JD_aws_senior_fde_zurich.txt
+80
View File
@@ -0,0 +1,80 @@
Skip to main content
Amazon Jobs home page
My career
Senior Forward Deployed Engineer, AWS Forward Deployed Engineering
Job ID: 10504263 | AWS EMEA SARL (Switzerland Branch)
Apply now
Description
AWS has formed a new Forward Deployed Engineer (FDE) team dedicated to embedding AI Engineers and Scientists directly inside strategic enterprise customer environments to help design, build, and deploy AI-powered production systems. We don’t build from a distance—we sit with customers, work in their infrastructure, and deliver production-grade AI solutions that transform how they operate. This initiative is scaling rapidly.
The Senior Forward Deployed Engineer (FDE) embeds directly inside a strategic enterprise customer to design, build, deploy, and run AI-powered production systems alongside the customer's engineering team. Senior FDEs write production-grade software and own outcomes end-to-end, from prototype through enterprise-scale deployment combining the technical rigor of an AWS SDE with the urgency and ownership required to deliver measurable customer business outcomes. They create reusable architectures, patterns, and engineering mechanisms that scale the entire FDE practice.
Key job responsibilities
• Embed within customer engineering teams. Understand the customer's business processes, technical architecture, and operational constraints. Translate ambiguous business problems into scalable AI-enabled production systems. Requires up to 30 – 50% travel.
• Build production AI applications. Design and develop production-grade software powering AI and agentic workflows — orchestration layers for multi-agent systems, retrieval pipelines, workflow automation, decision intelligence. Integrate foundation models, customer data sources, APIs, and existing applications into cohesive AI experiences. Optimize for latency, reliability, observability, cost, and security.
• Own production. Take systems from design through production rollout and operationalization. Troubleshoot production incidents across AI models, distributed systems, data pipelines, and application services. Put in place monitoring, evaluations, guardrails, and the operational mechanisms (testing, CI/CD, rollback, resiliency) that keep AI systems healthy at scale.
• Accelerate customer transformation. Identify opportunities to expand AI adoption across customer workflows. Codify reusable patterns, accelerators, and reference architectures that benefit future customers and feed back into AWS.
• Build and maintain internal AI tools and customer-facing assets and AI-enbled products to accelerate and scale the FDE motion.
• Route field signal into the closed-loop product feedback mechanism such as filing PFRs, escalating deployment-critical gaps, and feeding delivery learnings that become the next AWS primitives (as we already do with AWS Transform and Kiro).
• Raise the bar. Mentor engineers. Contribute to AWS engineering best practices for AI and forward deployment.
Basic Qualifications
- - 5+ years of non-internship professional software development experience
- - 5+ years of programming with at least one software programming language experience
- - 5+ years of leading design or architecture (design patterns, reliability and scaling) of new and existing systems experience
- - Experience as a mentor, tech lead or leading an engineering team
Preferred Qualifications
- Bachelor's degree in computer science or equivalent
- 5+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
- Experience building production AI or ML applications, including agentic workflows
- Strong written and verbal communication skills; comfort working in customer-facing environments
Amazon is an equal opportunities employer. We believe passionately that employing a diverse workforce is central to our success. We make recruiting decisions based on your experience and skills. We value your passion to discover, invent, simplify and build. Protecting your privacy and the security of your data is a longstanding top priority for Amazon. Please consult our Privacy Notice (https://www.amazon.jobs/en/privacy_page) to know more about how we collect, use and transfer the personal data of our candidates.
Our inclusive culture empowers Amazonians to deliver the best results for our customers. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, including support for the interview or onboarding process, please visit https://amazon.jobs/content/en/how-we-hire/accommodations for more information. If the country/region you’re applying in isn’t listed, please contact your Recruiting Partner.
Job details
CHE, ZH, Zurich
Software Development
Share this job
JOIN US ON
DOWNLOAD OUR APP
Find Careers
Job Categories
Teams
Locations
Military careers
Hourly
Working At Amazon
Culture
Benefits
Amazon Newsletter
Inclusive experiences
Leadership Principles
Help
FAQ
How We Hire
Application status
Accommodations
Legal disclosures and notices
English
Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status.
Privacy and Data
Impressum
Cookies
© 1996-2026, Amazon.com, Inc. or its affiliates
Notice of Right to Opt-Out Sharing
We use cookies and similar tools to provide and improve our site and your candidate experience. We also use operational cookies to make our site work. If you agree, we will set certain analytics cookies to help us improve our website as explained in our cookies notice, as well as set advertising cookies to display ads. Advertising cookies includes first and third-party cookies, which store or access standard device information such as a unique identifier. Third-parties use cookies to serve you with relevant advertisements and job recommendations based on your interactions with Amazon.Jobs and third-party sites. We will only set analytics or advertising cookies if you accept them. Select “Accept all” to consent or “Customize cookies” to make more detailed choices and learn more.
For more information about how we use cookies and information we collect, please navigate to your country’s cookie notice (if applicable) via our Privacy Notice and Data Policy page.
Accept all
Continue without accepting
Customize cookies
output/AWS_Senior_FDE_Zurich/session_aws_senior_fde_zurich.md
+133
View File
@@ -0,0 +1,133 @@
# Session: Amazon Web Services (AWS EMEA SARL, Switzerland Branch) — Senior Forward Deployed Engineer, AWS Forward Deployed Engineering
## JD Integrity
- File/source: `output/AWS_Senior_FDE_Zurich/JD_aws_senior_fde_zurich.txt` — amazon.jobs job ID **10504263**, `https://www.amazon.jobs/en/jobs/10504263`
- Retrieval method and date: Playwright headless scrape via `job_scout/.venv`, **2026-08-18**. The `amazon.jobs/*.json` endpoint returns HTTP 406, so the rendered page body was captured.
- Verbatim posting: **YES** — full visible posting body, not reconstructed.
- Posting status: **LIVE**, posted **2026-08-17** (one day before retrieval). Legal entity AWS EMEA SARL (Switzerland Branch), Zürich.
## Application Decision
- Audience profile: **International Tech** (US-tech conventions, role in Europe — config.md default)
- Evidence Fit: **69/100**
- Fit class: **Adjacent** (60–74)
- Hard gate: **PASS on minimum qualifications — but see the title-defining caveat below.** All four Basic Qualifications are Direct. No minimum qualification is a Gap.
- Channel Strength: **Weak** (cold application; no AWS contact recorded)
- Channel plan: Cold today. AWS Zurich + a brand-new org announced 2026-06-30 means recruiters are actively sourcing — a targeted note to the FDE recruiter is the highest-value warm action available.
- Cohort slot: Core 1/7, **Adjacent 1/2 → this would take the last Adjacent slot (2/2)**, Stretch 0/1
- Decision: **NO-GO — CLOSED 2026-08-18, user declined at the Phase 0 gate.**
69/100 is below the bar the user is willing to spend a package on, and the
title-defining Gap stands. **Cohort slot NOT consumed — Adjacent stays 1/2.**
The gate worked as designed: this was caught before any bullet was written.
### The caveat that must not be buried
`critique_framework.md` triggers a NO-GO flag when "a title-defining capability is not Direct."
**Forward deployment — embedding inside a strategic enterprise customer — is exactly that, and it is a Gap.**
The evidence base actively forbids the claim:
- `claims.json` **SW-4 forbidden:** `"customer-embedded delivery"`, `"strategic customer delivery"`, `"C-suite advisor"`
- `claims.json` **VZ-1 forbidden:** `"customer-embedded"`
- `claims.json` **global_forbidden_output_patterns:** `"customer-embedded delivery"`
- `application_strategy.md` lists **"strategic-account FDE" under Stretch targets**
- `historical_outputs.json` marks `output/Microsoft_ISE_Senior_SWE` unsafe for exactly this: *"customer-delivery framing exceeds verified evidence"*
So the single phrase that best describes this job is one we are never allowed to write about Dennis.
This is scored as Adjacent rather than No-Go because AWS's *published minimum bar is pure software
engineering* (see Requirements table) — they are hiring builders into a new org and transferring the
embedded model, not requiring prior consulting tenure. That is a real argument, not a rationalisation,
but it is an argument about AWS's intent, not about Dennis's evidence.
## Requirements
| # | Requirement | Required/preferred | Direct/Adjacent/Gap/Constraint | Canonical evidence | Gate? |
|---|---|---|---|---|---|
| B1 | 5+ years non-internship professional software development | Required | **Direct** | Continuous employment 2014-11 → present (~11.5 yrs) | Yes — PASS |
| B2 | 5+ years programming with at least one language | Required | **Direct** | Python, Java, C#, C++ across SWISSCOM/BOSCH/FRAUNHOFER/VIZRT (BS-2, VZ-1, FC-3) | Yes — PASS |
| B3 | 5+ years leading design or architecture (design patterns, reliability, scaling) of new **and existing** systems | Required | **Direct** | SW-2 Component Owner of business-critical Fulfillment ETL (operation, data quality, incidents, on-call); BS-3 Application Owner with SLOs; SW-3 Python apps on Kubernetes + GitLab CI/CD; SW-7 governed data products | Yes — PASS |
| B4 | Experience as a mentor, tech lead, or leading an engineering team | Required | **Adjacent** | Staff Engineer IV (Apr 2025→); BS-3 training and documentation; GN-1 trained teams on BDD/test automation; Leadership Cohort 2025. **No formal tech-lead title or direct reports.** | Yes — PASS (hedged) |
| P1 | Bachelor's in CS or equivalent | Preferred | **Direct** | M.Eng. Computer Aided Engineering (Software Design & Engineering), UniBw München | No |
| P2 | 5+ years full SDLC — coding standards, code review, source control, build, testing, operations | Preferred | **Direct** | VZ-2 automated A/V integration tests wired to CI/CD; FC-1 Jenkins CI/CD; SW-3 GitLab CI/CD; GN-1 test automation | No |
| P3 | Experience building production AI or ML applications, **including agentic workflows** | Preferred | **Gap (agentic) / Adjacent (production ML)** | BS-1 integrated containerized ML inference into 24/7 semiconductor production — genuine production ML. **But SW-8 forbids `"agent orchestration"`, `"built production LLM systems"`.** No RAG, no vector DB, no multi-agent framework anywhere in claims.json | No |
| P4 | Strong written/verbal communication; comfort in customer-facing environments | Preferred | **Adjacent** | SW-4 delivers data products with **internal B2B** stakeholders and product owners — explicitly *"not external consulting or strategic-account delivery"* | No |
| R1 | Embed within customer engineering teams; **up to 30–50% travel** | Responsibility (title-defining) | **Gap + Constraint** | No evidence; forbidden phrasing (SW-4, VZ-1, global). Travel itself is workable — user is travel-OK from a Bern base, no relocation | — |
| R2 | Build production AI apps: orchestration layers for multi-agent systems, retrieval pipelines, workflow automation | Responsibility | **Gap** | SW-8 forbids agent orchestration. Real adjacent: configured domain-grounded LLM assistants, LiteLLM gateway, custom GPTs, Copilot, **Kiro** | — |
| R3 | Own production — rollout, incidents across models/distributed systems/data pipelines, monitoring, CI/CD, rollback, resiliency | Responsibility | **Direct** | SW-2 (incidents, on-call, data quality), SW-3 (K8s + CI/CD), BS-3 (SLOs), BS-4 (ELK/Kafka anomaly detection + monitoring) | — |
| R4 | Codify reusable patterns, accelerators, reference architectures | Responsibility | **Adjacent** | SW-7 data-product modelling and documentation; BS-5 Spotfire C# extensions co-owned | — |
| R5 | Build internal AI tools to scale the FDE motion | Responsibility | **Adjacent** | SW-8 LiteLLM APIs, custom GPTs with fed domain knowledge | — |
| R6 | Route field signal into product feedback (PFRs) | Responsibility | **Adjacent** | SW-4 stakeholder translation; BS-3 vendor management | — |
| C1 | Location Zürich; work authorization | Constraint | **Direct** | German citizen (EU) + Swiss B permit — no sponsorship. Bern→Zürich ~1h by rail | — |
| C2 | English working language | Constraint | **Direct** | English fluent; German native (bonus for DACH accounts) | — |
**Tally:** Required 3 Direct + 1 Adjacent, **0 Gap** → minimum-qualification gate PASSES.
Responsibilities: 1 Direct, 3 Adjacent, **2 Gap** — one of which is title-defining.
## Evidence Fit — dimension table
| Dimension | Weight | Score | Reasoning |
|---|---:|---:|---|
| Required qualifications | 35 | **29** | All four Basic Quals Direct/near-Direct; B4 hedged (no formal lead title). Preferred P3 agentic is a Gap. |
| Core responsibilities | 25 | **11** | Only R3 (own production) is Direct. R1 embedding and R2 agentic are Gaps; R1 is title-defining. |
| Level and ownership | 15 | **9** | Staff Engineer IV + Component/Application Owner matches "Senior" IC scope well. No external-account authority. |
| Recency and depth | 10 | **8** | Current, substantial: AWS, Kubernetes, Python, production ownership all present-tense at Swisscom. |
| Domain/tool transfer | 10 | **7** | **AWS is the ecosystem match** (contrast: claims.json marks GCP `output: forbidden`). Bedrock/AgentCore unused but adjacent to LiteLLM. Kiro genuinely used. |
| Practical constraints | 5 | **5** | Zürich commutable, EU citizen + B permit, English, travel acceptable. |
| **Total** | **100** | **69** | **Adjacent** |
## Competitive Read
- **Obvious-fit candidate:** a consulting-side senior engineer from AWS ProServe, Palantir, Accenture Applied Intelligence or a Big-4 AI practice who has already shipped LLM/agentic systems *inside* client environments, and who is used to 50% travel and client politics.
- **Dennis's advantage:** genuine **production ownership** — on-call, incidents, SLOs, data quality on business-critical pipelines. Many FDE applicants are strong at prototypes and weak at running things. He also brings AWS-native depth, native German for DACH accounts, and hands-on Kiro use (AWS's own agentic IDE, named in the JD).
- **Their advantage:** they have actually done the title-defining thing — embedded delivery against strategic accounts — and they have the agentic/RAG production track record listed in the preferred quals.
- **Level/scope comparison:** Level matches (Senior IC, not people-management). The gap is *function*, not seniority. That is the honest read.
## Company Context
- **The org is 7 weeks old.** AWS announced Forward Deployed Engineering on **2026-06-30** with a **$1bn** commitment, planning to deploy *thousands* of engineers. Model: small teams of ~5–6 embedded at customer sites for ~45-day engagements. Explicitly positioned as **not traditional consulting** — the stated goal is transferring expertise so customers operate the systems themselves. The JD's "This initiative is scaling rapidly" is literal.
- **Why that matters for Dennis:** a brand-new org hiring at volume screens differently from a mature consulting practice. They need builders who can run production, and they are training the embedded motion. This is the strongest structural argument for an Adjacent application.
- **Kiro** — the JD says "as we already do with AWS Transform and Kiro." Kiro is AWS's spec-driven agentic IDE, GA March 2026, successor to Amazon Q Developer, routing Claude Sonnet + Amazon Nova over Bedrock. **Dennis uses Kiro** (config.md verified GenAI toolchain). This is a rare, first-party, verifiable hook — not a vocabulary match.
- **AWS Zurich** office is operational with active hiring (Delivery Consultant, Practice Manager, Solutions Manager roles alongside this one).
## Framing Strategy
- **Professional identity:** Staff-level engineer who **builds and runs production systems on AWS** — pipelines, data products, containerized services — and has integrated ML inference into a 24/7 industrial environment.
- **Lead narrative:** *"I don't hand over prototypes; I own what happens at 3am."* Production ownership is the honest differentiator against a prototype-heavy FDE applicant pool.
- **Strongest proof points:** SW-2 Component Owner (incidents/on-call/data quality) · BS-1 containerized ML inference into 24/7 semiconductor production · SW-3 Python on Kubernetes + GitLab CI/CD · SW-7 governed data products on AWS · BS-4 ELK/Kafka anomaly detection.
- **Honest adjacent bridges:**
- AI work → frame as **integration and enablement** (configured domain-grounded assistants, LiteLLM gateway, Kiro, Copilot), never as agent orchestration.
- Customer-facing → frame as **internal B2B stakeholder delivery and vendor/Application ownership** (SW-4, BS-3), never as embedded or strategic-account delivery.
- **Explicit gaps — do not paper over:** no customer-embedded delivery; no multi-agent/RAG production systems; no Bedrock/AgentCore.
- **Downplay:** Security Champion (JD has no security requirement — CLAUDE.md default is OMIT); semiconductor domain depth beyond the ML-inference story.
- **User directives:** none given for this JD.
## Critique Context
- **Reviewer persona:** an AWS FDE hiring manager staffing a brand-new org at speed — pragmatic, bar-raiser-trained, scanning for "can this person build *and* operate under customer pressure."
- **Amazon-specific:** Leadership Principles matter. *Ownership*, *Bias for Action*, *Dive Deep*, *Deliver Results* are the ones his evidence genuinely supports. Amazon interviews are LP-heavy — worth noting for interview prep, not for resume prose.
- **Domain vocabulary:** forward deployed, production-grade, end-to-end ownership, foundation models, Bedrock, agentic workflows, evaluations, guardrails, observability, latency/cost optimization, reference architectures, PFR.
- **Likely first technical challenge:** *"Walk me through an AI system you took to production."* The honest answer is BS-1 (containerized ML inference, 24/7 fab) — classical ML, not LLM. Prepare that answer deliberately; do not let it drift toward implying LLM production ownership.
## Cover Letter Decision
- **YES**
- **Reason:** The resume cannot, on its own, explain why a Swisscom platform engineer is applying to an embedded customer-delivery role. That gap is visible on page one and will otherwise be resolved against him. A letter is the only place to address it directly and turn production ownership into the argument.
- **Information it adds beyond resume:** (1) the motivation for moving from internal platform work into forward-deployed delivery, stated plainly rather than disguised; (2) **hands-on Kiro use** — first-party AWS tooling named in the JD; (3) native German for DACH accounts; (4) travel/mobility willingness from a stable Swiss base.
- **Verified hook sources:** JD text (Kiro, AWS Transform, PFR mechanism); AWS FDE launch 2026-06-30 ($1bn, ~45-day embedded engagements); config.md verified GenAI toolchain (Kiro, Copilot, LiteLLM, custom GPTs).
- **Forbidden in the letter:** any phrasing implying prior customer-embedded or strategic-account delivery; any agent-orchestration claim.
## Resume Plan
- Summary: 2–3 lines — production ownership + AWS + ML-into-production
- Skills: 4–6 evidence-backed lines; AWS prominent; **no Bedrock, no LangGraph/CrewAI/ADK, no vector DB**
- Swisscom: SW-2, SW-3, SW-7, SW-1, SW-8 (integration framing only)
- Bosch: BS-1 (lead), BS-3, BS-4
- Earlier experience: VZ-1 or VZ-2 (distributed systems / CI/CD), FC-3 optional
- Total bullets: 11–14
- Impact evidence still needed: **all Swisscom metrics are `unverified`** — this remains the standing weakness across every package.
## Output Files
- JD (verbatim): `output/AWS_Senior_FDE_Zurich/JD_aws_senior_fde_zurich.txt`
- Resume: pending Phase 2
- Cover letter: pending (decision YES)
- Critique: pending
## Status
- Fit gate: **CLOSED — NO-GO 2026-08-18** (Phase 0 completed, then declined)
- Original gate result: **Phase 0 DONE 2026-08-18** — Adjacent, Evidence Fit 69/100, minimum-qual gate PASS, title-defining Gap disclosed, Channel Weak
- Phase 0: DONE
- Phase 1: **NOT STARTED — will not start.**
- Resume: not generated
- Cover Letter: not generated (CL decision was YES, but moot)
- Critique: n/a
output/AkerBP_Data_Product_Architect/session_akerbp_data_product_architect.md
+1 -1
View File
@@ -152,6 +152,6 @@ All hooks verified against first-party or canonical sources. **No unverified cla
- Critique: **CURRENT** — Evidence Fit 79/100 Adjacent, Document Quality 93/100, hard gate PASS, Channel Weak. 3 Tier 1 findings applied (architecture vocabulary, Languages line, catalogue/adoption terms). See `critique_akerbp_data_product_architect.md`.
- Finalized: **YES 2026-07-29** — `Dennis_Thiessen_Resume.pdf`, `Dennis_Thiessen_Cover_Letter.pdf`; cohort entry added (evidence-first-2026-01, adjacent 1/2)
- Pre-submission actions outstanding: **attach diploma/grade transcript** (required by posting); optional cold email to Per Olav Marthinsen
- Application/outcome: **SUBMITTED 2026-07-29** (4 days before the 2026-08-02 deadline). Cohort outcome set to `applied`. Awaiting response.
- Application/outcome: **SUBMITTED 2026-07-29** (4 days before the 2026-08-02 deadline). **CLOSED — REJECTED 2026-08-27, no interview, ~28 days to disposition.** Cohort outcome set to `rejected_no_interview` (evidence-first-2026-01, Adjacent slot stays consumed — a rejection does not free it); decision log updated. Cold channel: the optional follow-up email to Per Olav Marthinsen was never actioned, so this went in as a pure cold submit. Package was validator PASS with no Tier 1 truth findings — **this outcome does not implicate document quality on its own**, and per `application_strategy.md` §38/§74 no positioning change follows from a single rejection. Note: `CLAUDE.md` and the decision log record the submission as 2026-07-30 — a one-day inconsistency with this file, left unreconciled.
- Next: `/make-cl output/AkerBP_Data_Product_Architect/session_akerbp_data_product_architect.md`
- Then: `/critique output/AkerBP_Data_Product_Architect/session_akerbp_data_product_architect.md`
output/CitadelSecurities_PlatformEngineer/JD_CitadelSecurities_PlatformEngineer.txt
+59
View File
@@ -0,0 +1,59 @@
SOURCE URL: https://www.citadelsecurities.com/careers/details/platform-engineer/
RETRIEVED: 2026-08-25 via job_scout/.venv Playwright (chromium, headless)
METHOD NOTE: WebFetch returns HTTP 403 on this host; Playwright required.
VERBATIM: yes - full visible posting body, unedited except removal of site
navigation chrome (menu, footer, cookie/legal links).
========================================================================
Platform Engineer
New York, Miami, Zurich
Job Description
Role Overview:
Citadel Securities is seeking an exceptional Platform Engineer to join our Research Platform Engineering team in Miami, Zurich or New York. Our Research Platform teams are at the forefront of designing and developing the distributed systems, data platforms, and research infrastructure that power large-scale simulations, analytics, and model development across the firm. These systems are critical in allowing researchers to transform data into actionable insights and alpha-generating trading strategies.
Opportunities may be available from time to time in any location in which the business is based for suitable candidates. If you are interested in a career with Citadel, please share your details and we will contact you if there is a vacancy available.
Responsibilities:
Design and build distributed platforms and services that support quantitative research, simulations, AI-powered applications, and large-scale analytics workflows
Develop scalable platform services and data systems for ingestion, transformation, storage, and lifecycle management of large datasets
Build and maintain backend services, APIs, and SDKs that improve researcher productivity and self-service capabilities
Improve performance, reliability, and scalability of critical research systems operating in high-throughput environments
Collaborate closely with researchers and engineers to translate research requirements into production-read systems
Qualifications:
Bachelor’s degree in Computer Science or a related field
3+ years of professional software engineering experience
Strong programming skills in Python, Go, or C++, or similar systems-oriented languages
Experience building and operating distributed systems in production environments
Experience designing scalable backend services, APIs, and data-intensive applications
Strong understanding of distributed systems, data-intensive applications, and system design fundamentals
Experience with cloud platforms (AWS, GCP, Azure) and modern infrastructure technologies
Experience with technologies such as Kafka, Kubernetes, Spark, Airflow, distributed databases, or similar systems is a plus
In accordance with applicable law, the base salary range for this role is $175,000 to $350,000.
In addition, the employee who fills this role will be eligible to participate in a discretionary incentive compensation program, as well as a wide array of benefit programs, such as medical and life insurance, retirement and tax-free savings plans, and access to other healthcare programs.
About Citadel Securities
Citadel Securities is a technology-driven, next-generation global market maker. We provide institutional and retail investors with world-class liquidity, competitive pricing and seamless front-to-back execution in a broad array of financial products. Our teams of engineers, traders and researchers harness leading-edge quantitative research and the accelerating power of compute, machine learning and AI to power our analytics and tackle the market’s and our clients’ most critical challenges. Together, we are forging the future of capital markets. For more information, visit citadelsecurities.com.
About Citadel Securities
Citadel Securities is a technology-driven, next-generation global market maker. We provide institutional and retail investors with world-class liquidity, competitive pricing and seamless front-to-back execution in a broad array of financial products. Our teams of engineers, traders and researchers harness leading-edge quantitative research and the accelerating power of compute, machine learning and AI to power our analytics and tackle the market’s and our clients’ most critical challenges. Together, we are forging the future of capital markets. For more information, visit citadelsecurities.com.
Start Application
Citadel Securities
Counterparty Login
Who We Are
What We Do
News & Insights
Careers
Privacy
Terms of Use
Fraud Notice
Disclosures
Copyright © Citadel Enterprise Americas LLC or one of its affiliates. All rights reserved.
Citadel Securities is an equal opportunity employer. We provide all individuals consideration for employment and advancement opportunities without regard to race, religion, color, gender, pregnancy, national origin, age, disability, military or veteran status, sexual orientation, genetic information and any other classification protected by applicable federal, state and local laws.
output/CitadelSecurities_PlatformEngineer/critique_citadel_securities_platform_engineer.md
+190
View File
@@ -0,0 +1,190 @@
# Critique — Citadel Securities, Platform Engineer (Research Platform Engineering)
**Date:** 2026-08-26 · **Package:** 2-page English resume, 13 bullets · **Cover letter:** not yet generated
**Framework:** `critique_framework.md` §9 (6-dimension Document Quality). The SKILL.md 8-dimension
table with "Publications 10%" is a stale CV-era scheme and was not used — same conflict class as its
"all bullets 2L" line, which `resume_reference.md` §7 superseded on 2026-07-27.
---
## 1. Fit Verdict and Hard Gates
**Qualifications gate: PASS.** All seven minimum quals are Direct or Adjacent; none is a Gap; nothing
required is `forbidden` in `claims.json`.
**Practical gate: FAIL — UNRESOLVED.** The Zurich working model is unknown. Ken Griffin is on record
that Citadel returned to the office five days a week and calls it his most important leadership
decision. If that holds for Zurich, the seat needs ~2–2.5h daily commuting from Bern or relocation,
against a standing Bern-based / 2–3 day hybrid constraint. `critique_framework.md` §1 treats an
unresolved location constraint as a NO-GO trigger.
> **This document is finished. The application is not cleared.** Do not submit until the working
> model is answered. The resume was built at user direction with this open.
## 2. Evidence Fit — 71/100 (Adjacent, mid-band) — unchanged
| Dimension | Weight | Score |
|---|---:|---:|
| Required qualifications | 35 | 29 |
| Core responsibilities | 25 | 16 |
| Level and ownership | 15 | 10 |
| Recency and depth | 10 | 9 |
| Domain/tool transfer | 10 | 6 |
| Practical constraints | 5 | **2** |
Evidence Fit describes the pairing, not the document; writing cannot move it. R1 (design and build
the research platform) stays Adjacent regardless of prose quality.
## 3. Document Quality — 86/100
| Dimension | Weight | Score | Note |
|---|---:|---:|---|
| Truth and provenance | 25 | **23** | No fabrication, no invented metric, no forbidden tool. Scope discipline held in every bullet. PP-1 labelled three ways with zero performance figures. One borderline item in the headline (below). |
| Information hierarchy | 20 | **18** | Conventional employer/title/date hierarchy; strongest evidence on page 1; Personal Project correctly placed after experience and before education. Page 1 carries visible bottom whitespace from the deliberate page break. |
| Bullet evidence and impact | 20 | **16** | Specific and well-scoped, but several bullets describe duties rather than demonstrate outcome (SW-2, BS-3), and the document contains **zero verified metrics** — a standing profile issue, not a writing failure. |
| Relevance and terminology | 15 | **10** | **The weakest dimension. See Tier 1.** 19/37 JD terms present (51%). |
| Skills evidence | 10 | **10** | Six lines, every entry canonical, C++ and the trading domain both correctly caveated, no self-ratings, nothing listed merely because the JD says it. |
| Mechanics/readability | 10 | **9** | Compiles to exactly 2 pages, PDF clean. −1 for cadence per §7a cap. |
| **Total** | **100** | **86** | |
Truth/provenance 23/25 is far above the 8/10 automatic-failure line.
## 4. Channel Strength — **Weak**
Cold application, no named contact on the posting, no known connection. This would be the fifth
consecutive cold-channel application. Tailoring does not upgrade a cold channel.
## 5. Competitive Read
**The obvious-fit competitor:** an ex-FAANG infrastructure or distributed-systems engineer, likely
already in Zurich, who has *designed* platform components rather than owned pipelines on one. The
Zurich office is visibly staffed from Google and ETH.
| | Dennis | Competitor |
|---|---|---|
| Direct function | Owns pipelines and data products **on** a platform | Builds the platform itself |
| Tool ecosystem | **Advantage** — exact stack match, production-current | Likely comparable or narrower |
| Domain | Self-directed only (PP-1, PP-2) | Usually none either — **near-parity** |
| Level/scope | Staff + Component Owner, above the stated bar | Varies |
| Access | Cold | Often referred |
**Dennis's real advantage:** production ownership with on-call — he has carried the pager for
revenue-adjacent systems, which most platform candidates have not. Plus genuine domain interest.
**Competitor's real advantage:** they have done R1, the title-defining half he has not.
## 6. Reader Sequence
- **Parser/ATS:** extracts cleanly. Standard LaTeX, conventional headers, no tables or columns to
break parsing. Term coverage 51% — passable, improvable (Tier 1).
- **Recruiter glance:** **Forward.** Staff title, Zurich-adjacent location, and the exact stack in the
headline and first skills line.
- **Hiring manager:** **Maybe.** R2 evidence is unmistakable; R1 evidence is absent and they will
notice. The Personal Project is the line most likely to earn a conversation.
- **Technical reviewer:** credible. **Likely first challenge: "Which parts of the platform did you
design, versus consume?"** The document does not pre-empt this, and it cannot honestly — but the
candidate must have a crisp scoped answer ready.
## 7. Claim Audit
| Claim | Canonical | Direct/Adjacent | Safe? | Fix |
|---|---|---|---|---|
| Summary: builds/runs pipelines + governed data products | SW-2, SW-3, SW-7 | Direct | ✅ | — |
| Summary: Component Ownership incl. on-call | SW-2 | Direct | ✅ | — |
| Summary: Bosch query-performance tuning | BS-2 | Direct | ✅ | — |
| Summary: containerized ML inference into operating fab | BS-1 | Direct | ✅ | — |
| **Headline: "Production Data Platforms on AWS"** | SW-1/SW-3/SW-7 | Adjacent | ⚠️ | **Tier 2** — no ownership verb, so not a violation, but it is the most prominent line in the document and leans toward the platform framing R1 is a gap on |
| Data products "within Swisscom's company-wide Data Mesh" | SW-7 | Direct | ✅ | Correctly scoped — the object carries "company-wide", the verb does not claim it |
| "Migrated owned … pipelines … contributing to the wider company migration programme" | SW-1 | Direct | ✅ | Correctly scoped and hedged |
| Component Owner, Fulfillment ETL | SW-2 | Direct | ✅ | — |
| Bosch data services + query tuning + fab data types | BS-2 | Direct | ✅ | — |
| ELK/Kafka anomaly-detection **proof of concept** | BS-4 | Direct | ✅ | "proof of concept" retained — good |
| Application Owner, SLOs, vendors, training | BS-3 | Direct | ✅ | — |
| Vizrt: "Contributed Python and C++ … distributed video-transcoding backend" | VZ-1 | Direct | ✅ | Hedged verb correct; no broadcaster names |
| PP-1 Personal Project block | PP-1 | Direct | ✅ | Labelled 3×; **no PnL/Sharpe/return/backtest figure** — compliant |
| PP-2 as "Supporting coursework" | PP-2 | Direct | ✅ | Not presented as degree or credential |
| Trading skills line, "personal project and coursework only, no professional experience" | PP-1/PP-2 skills | Direct | ✅ | Exactly the caveat `claims.json` requires |
`validate_resume_system.py --document` → **PASS** (1 cadence warning). **No Tier 1 truth findings.**
## 8. Tiered Improvements
### Tier 1 (≥1 pt) — Relevance and terminology
**1. The strongest matching responsibility is the least vocabularised.** R2 — *"ingestion,
transformation, storage, and lifecycle management of large datasets"* — is the one core
responsibility scored **Direct**, and it is the closest thing to a plain description of SW-1/SW-7.
Yet **"ingestion", "transformation", "storage" and "lifecycle" are all absent** from the resume body.
The document proves R2 in substance while missing it in the JD's own words, which is exactly what an
ATS and a skimming reviewer key on. All four are honestly available: onboarding source systems *is*
ingestion; ETL *is* transformation; S3/Redshift/Iceberg *are* storage. **Worth ~2 pts.**
**2. "distributed systems" never appears as a phrase.** Q4 is a *required* qualification and Phase 0
scored it Direct — a production Kafka/Kubernetes/Spark/Airflow estate is a distributed system. The
resume says "Distributed data systems" (skills header) and "distributed video-transcoding backend"
(Vizrt, 2018), so the nearest exact match sits on an eight-year-old bullet rather than current work.
**Worth ~1 pt.** Fix by naming the current estate as distributed — without claiming to have *designed*
distributed systems, which remains a Gap.
**3. "distributed databases" is claimed as a Direct preferred hit but never written.** Phase 0 called
all five preferred technologies Direct. Four are named explicitly; the fifth appears only as product
names (Teradata, Redshift, Athena/Iceberg). Teradata is MPP and Redshift is distributed, so the term
is honest — it is simply missing. **Worth ~1 pt.**
### Tier 2 (0.3–0.9 pt)
4. **Tighten the headline.** "Production Data Platforms on AWS" → something like "Production Data
Pipelines & Data Products on AWS". More precisely his scope, still JD-relevant, and removes the
one line a technical reviewer could open by challenging.
5. **"infrastructure"** is absent; CloudFormation, Kubernetes and Ansible support it honestly.
6. **Page 1 bottom whitespace.** Do not pad (§9 forbids it), but Tier 1 fixes 1–3 will naturally
absorb some of it.
### Tier 3 (<0.3)
7. "system design" — supportable via iSAQB CPSA-F but risky next to an R1 gap; probably leave.
8. Cadence 58% vs the ~50% guide. **Do not chase this.** The residual flags are legitimate technology
enumerations, and the §7a checker cannot distinguish those from rhetorical tricolons.
### Deliberately NOT to be added
- **"scalable"** — an unverified scale claim. `ai_fingerprint_rules.md` forbids converting scale into
personal impact, and `claims.json` marks all metrics unverified.
- **SDK, high-throughput, data-intensive, simulation, model development, self-service** — zero
canonical evidence. Their absence is correct, not a defect.
## 9. Cover Letter
**Not generated.** Session decision is **YES**, and that still stands: the resume can carry PP-1 only
as a compact block, and cannot explain *why* a Swisscom data engineer credibly wants a research-platform
seat at a market maker. Against an ex-FAANG field that motivation is the main differentiator and it
needs prose.
**Does the resume stand alone?** Partly. It earns a Forward from a recruiter and a Maybe from a hiring
manager on stack alignment. What it cannot do alone is answer "why us, and why would you leave a Staff
seat for this" — which, given a cold channel, is the question that decides the screen.
## 10. Mechanical Verification
- ✅ `validate_resume_system.py --document` → PASS (1 cadence WARN)
- ✅ Compiles, MiKTeX pdflatex, **exactly 2 pages**
- ✅ Rendered PDF inspected: no clipping, overlap, orphaned heading or bad page break
- ✅ Umlauts, `C\#`, `C++`, en-dashes all render correctly
- ✅ Email matches `config.md` (`dennis@thiessen.io`)
- ✅ No LOC counts, no test counts, no Security Champion, no forbidden tools
- ✅ FIXED sections (Education, header block) untouched
- ⚠️ `char_count.py` output unusable for bullets 9–12: its parser does not respect `rSubsection`
boundaries and bleeds following-section text into each position's last bullet. Pre-existing tool
limitation; rendered PDF used as authority instead.
---
## Verdict
**Document Quality 86/100 · Evidence Fit 71/100 · Channel Weak · Practical gate UNRESOLVED.**
The document is honest, clean and well-scoped, with no truth findings. Its one real weakness is that
it under-uses the JD's own language for the very responsibility it matches best — three Tier 1 fixes,
all honestly available, worth roughly +4.
None of that changes the decision. **Document quality is not the constraint here; the working model
and the cold channel are.** Do not treat 86 as submit-ready.
output/CitadelSecurities_PlatformEngineer/e2e_citadel_securities_platform_engineer_cover_letter.tex
+57
View File
@@ -0,0 +1,57 @@
% Citadel Securities — Platform Engineer, Research Platform Engineering (Zurich)
% Cover Letter Decision: YES — domain transition needs explaining; the resume can carry PP-1
% only as a compact block and cannot show why a Swisscom data engineer wants this seat.
% Industry format per cl_reference.md: 3 short paragraphs, 200--300 words.
% Hooks: JD text only (first-party, scraped verbatim). No named executives — cl_reference.md
% says omit an unnecessary hook rather than spend words proving company familiarity.
% Every claim traceable to the resume. PP-1: no performance figure, labelled personal.
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}
\usepackage[a4paper,left=0.85in,right=0.85in,top=0.7in,bottom=0.7in]{geometry}
\usepackage{parskip}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\begin{document}
{\Large\bfseries Dennis Thiessen, M.Eng.}\par
Bern, Switzerland $\vert$
\href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$
\href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}
\vspace{1.5em}
Research Platform Engineering\\
Citadel Securities\\
Zurich, Switzerland
\vspace{1em}
26 August 2026
\vspace{1em}
\textbf{Application for Platform Engineer, Research Platform Engineering}
\vspace{0.8em}
Dear Hiring Team,
The posting describes data systems for ingestion, transformation, storage and lifecycle management of large datasets, built so that researchers can work without fighting their infrastructure. That is my day job. At Swisscom I am Component Owner for business-critical Fulfillment ETL -- a distributed system across Oracle, Kafka, Python and Teradata -- and I build the governed data products that analytics and AI teams consume. I carry the data quality and the on-call pager for that estate, which is a different relationship to a system than having built it and moved on.
What draws me to research platform work specifically is that I already build this kind of system for myself. Outside work I built a self-hosted equity signal platform: it ingests daily prices, fundamentals and sentiment, runs a long-only cross-sectional momentum book through scheduled scan and backtest pipelines, and surfaces the gated setups to a dashboard. It is a personal system, not a production service, and I make no claims for its results. But it is the reason I understand what a researcher actually needs from a platform -- fast iteration, reproducible runs, and trust that the data underneath is right -- rather than knowing it only as a requirements document.
I should be straightforward about the shape of my experience: I have built and operated data products and pipelines on a platform, not authored a research platform at firm scale. That step is the reason the role interests me, and the operational judgement I would bring to it is real. I am based in Bern, hold EU citizenship and a Swiss B permit, and need no sponsorship.
\vspace{0.8em}
I would welcome the chance to talk.
\vspace{1em}
Kind regards,
\vspace{1.5em}
Dennis Thiessen
\end{document}
output/CitadelSecurities_PlatformEngineer/e2e_citadel_securities_platform_engineer_resume.tex
+92
View File
@@ -0,0 +1,92 @@
% Citadel Securities — Platform Engineer, Research Platform Engineering (New York, Miami, Zurich)
% Profile: International Tech, English, 2 pages. Generated 2026-08-26.
% Generated from resume_builder/canonical/claims.json + normalized experience files.
% Several tools named in the JD are non-canonical and are deliberately absent from this
% document; they are listed and bridged in the session file's Gap Assessment. Do not
% reintroduce them here, not even in a comment -- the validator scans raw file text.
% JD qualifications 3 and 7 are disjunctive, so Python and AWS satisfy them alone.
% PP-1 appears only under an explicitly labelled Personal Project heading; claims.json
% forbids calling it distributed/production/multi-user or quoting ANY performance figure.
\documentclass{resume}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}
\usepackage[a4paper,left=0.65in,right=0.65in,top=0.55in,bottom=0.55in]{geometry}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\usepackage[version=4]{mhchem}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[R]{\small \thepage/\pageref{LastPage}}
\name{Dennis Thiessen, M.Eng.}
\headline{Staff Data Engineer $\vert$ Production Data Platforms on AWS $\vert$ Kafka $\cdot$ Kubernetes $\cdot$ Spark}
\contactline{Bern, Switzerland $\vert$ \href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$ \href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}}
\begin{document}
\begin{rSection}{Summary}
Staff data engineer who builds and runs the pipelines and governed data products that analytics and AI consumers depend on, on AWS and Kubernetes at Swisscom, with Component Ownership covering data quality, incidents and on-call duty. Earlier work at Bosch tuned query performance over Oracle and Hadoop/Impala for semiconductor analysis teams and moved containerized ML inference into a continuously operating fab.
\end{rSection}
\begin{rSection}{Technical Skills}
\skillline{Programming}{Python, SQL, PySpark; Java and C\# (professional, historical); C++ (limited historical use)}
\skillline{Distributed systems}{Apache Kafka, Apache Airflow, Spark/PySpark, distributed databases (Teradata, Redshift, Athena/Iceberg), ETL/ELT, data modelling, query performance tuning, Oracle, Hadoop/Impala}
\skillline{Cloud and platform}{AWS (S3, Glue, Athena/Iceberg, Redshift, Lambda, Step Functions), CloudFormation, Kubernetes, Docker, GitLab CI/CD, Ansible}
\skillline{Reliability and operations}{Production operation and on-call, SLOs, incident and root-cause analysis; Grafana, Prometheus, ELK (proof of concept)}
\skillline{Trading domain (self-directed)}{Systematic/quantitative trading concepts, factor construction, backtesting -- personal project and coursework only, no professional experience}
\skillline{Certifications}{AWS Solutions Architect -- Associate; Data Engineering with AWS; iSAQB CPSA-F; ITIL Foundation}
\end{rSection}
\begin{rSection}{Professional Experience}
\begin{rSubsection}{Swisscom (Schweiz) AG}{Oct 2023 -- Present}{Staff Data, Analytics \& AI Engineer (promoted from Senior, Apr 2025)}{Bern, Switzerland}
\item Build and model governed data products within Swisscom's company-wide Data Mesh, from source-system ingestion through to the metadata and lineage in Atlassian Compass that keep each product discoverable across its lifecycle.
\item Component Owner for business-critical Fulfillment ETL -- a distributed system spanning Oracle, Kafka, Python and Teradata -- accountable for its data quality, governance and on-call operation.
\item Migrated owned Fulfillment and Product Analysis pipelines onto Swisscom's AWS platform -- transformation and storage on Glue, Athena with Apache Iceberg and Redshift, orchestration in Airflow -- contributing to the wider company migration programme.
\item Develop and operate Python data applications on Kubernetes, delivered through GitLab CI/CD from automated test to production support.
\item Translate stakeholder and product-owner requirements into delivered data products; run root-cause analysis in 2nd- and 3rd-level support.
\end{rSubsection}
\end{rSection}
\newpage
\begin{rSection}{Professional Experience (continued)}
\begin{rSubsection}{Robert Bosch Semiconductor Manufacturing Dresden GmbH}{Feb 2020 -- Dec 2022}{Senior Engineer, Data Analysis (Data Engineering)}{Dresden, Germany}
\item Developed data services in Python, Java and C\# over Oracle and Hadoop/Impala, tuning query performance for analysis teams working with semiconductor defect records, wafer inspection images and process-control electrical parameters.
\item Integrated containerized ML inference with Docker, Kubernetes and Ansible into a continuously operating semiconductor fab, replacing manual image-based defect classification.
\item Built an anomaly-detection proof of concept on Elasticsearch, Logstash, Kibana and Kafka, with Grafana and Prometheus monitoring for continuously running manufacturing systems.
\item Served as Application Owner for analytics applications and their upstream pipelines, defining SLOs and coordinating vendors, user training and documentation.
\end{rSubsection}
\begin{rSubsection}{Fraunhofer CML}{Sep 2018 -- Oct 2019}{Research Software Engineer}{Hamburg, Germany}
\item Developed containerized microservices for the MISSION maritime research data-exchange platform.
\end{rSubsection}
\begin{rSubsection}{Vizrt}{Jul 2017 -- May 2018}{Test Automation / DevOps Engineer}{Bergen, Norway}
\item Contributed Python and C++ engineering to a distributed video-transcoding backend.
\end{rSubsection}
\begin{rSubsection}{Generali Deutschland Informatik Services GmbH}{May 2015 -- Jun 2017}{IT Consultant}{Hamburg, Germany}
\item Introduced BDD test automation through a proof of concept and held technical responsibility for the suite and its Jenkins jobs.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Personal Project}
\compactentry{Equity Signal Platform -- self-hosted, single-user}{Personal project}
Ingests daily US-equity prices, fundamentals and LLM-assisted sentiment; runs one long-only cross-sectional momentum book (residual 12-1 momentum, ATR stop and trail, at most 15 concurrent positions) through scheduled scan and backtest pipelines; surfaces the gated setups in a web dashboard and Telegram alerts. Built end to end as a personal system, not a production or multi-user service. Supporting coursework: Udacity \textit{AI for Trading} nanodegree (factor construction, alpha signals, backtesting).
\end{rSection}
\begin{rSection}{Education}
\compactentry{M.Eng. Computer Aided Engineering (Software Design \& Engineering), Universität der Bundeswehr München}{Apr 2012 -- Oct 2013}
Thesis at Tongji University, Shanghai: \textit{Development of a Web-Based Remote Fault Diagnosis System}; grade 1.0.
\compactentry{B.Eng. Information and Telecommunication Technologies, Universität der Bundeswehr München}{Oct 2009 -- Oct 2012}
\end{rSection}
\end{document}
output/CitadelSecurities_PlatformEngineer/resume.cls
+59
View File
@@ -0,0 +1,59 @@
\ProvidesClass{resume}[2026/07/27 Evidence-first professional resume]
\LoadClass[11pt,a4paper]{article}
\RequirePackage{enumitem}
\RequirePackage{ifthen}
\RequirePackage{lastpage}
\RequirePackage{parskip}
\RequirePackage{needspace}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\def\@resumename{}
\def\@resumeheadline{}
\def\@resumecontact{}
\newcommand{\name}[1]{\def\@resumename{#1}}
\newcommand{\headline}[1]{\def\@resumeheadline{#1}}
\newcommand{\contactline}[1]{\def\@resumecontact{#1}}
\newcommand{\printresumeheader}{%
{\LARGE\bfseries \@resumename}\par
\vspace{0.15em}
\ifthenelse{\equal{\@resumeheadline}{}}{}{\@resumeheadline\par}
\vspace{0.1em}
{\small \@resumecontact}\par
\vspace{0.45em}
}
\AtBeginDocument{\printresumeheader}
\newenvironment{rSection}[1]{%
\vspace{0.35em}
{\bfseries #1}\par
\vspace{0.1em}\hrule\vspace{0.35em}
}{%
\vspace{0.2em}
}
% Arguments: employer, dates, formal title, location.
\newenvironment{rSubsection}[4]{%
\Needspace{6\baselineskip}%
{\bfseries #1}\hfill{\small #2}\par
{\itshape #3}\hfill{\itshape #4}\par
\vspace{0.1em}
\begin{itemize}[leftmargin=1.25em,label=\textbullet,itemsep=0.18em,topsep=0.15em,parsep=0pt,partopsep=0pt]
}{%
\end{itemize}
\vspace{0.25em}
}
\newcommand{\skillline}[2]{%
\textbf{#1:} #2\par
\vspace{0.08em}
}
\newcommand{\compactentry}[2]{%
\textbf{#1}\hfill #2\par
}
output/CitadelSecurities_PlatformEngineer/session_citadel_securities_platform_engineer.md
+395
View File
@@ -0,0 +1,395 @@
# Session: Citadel Securities — Platform Engineer, Research Platform Engineering
## JD Integrity
- File/source: `JD_CitadelSecurities_PlatformEngineer.txt` (this folder) — copied from `JDs/JD_CitadelSecurities_PlatformEngineer.txt`
- Source URL: https://www.citadelsecurities.com/careers/details/platform-engineer/
- Retrieval method and date: Playwright headless (chromium) via `job_scout/.venv`, 2026-08-25. **WebFetch returns HTTP 403 on this host** — Cloudflare-protected; do not use WebFetch here.
- Verbatim posting: YES — full visible posting body, unedited except removal of site navigation chrome.
- Posting status: **LIVE as of 2026-08-25.** Verified independently of the detail page: the role appears in the site's Zurich location filter, which returns 10 open Zurich roles. Board total 81 open roles.
- Locations: **New York, Miami, Zurich** (JD body: "in Miami, Zurich or New York")
- Team: Research Platform Engineering
- **Evergreen caveat:** the posting carries standing-req boilerplate — *"Opportunities may be available from time to time in any location in which the business is based for suitable candidates… we will contact you if there is a vacancy available."* Most likely a continuously-open multi-location req. Less deadline pressure; higher chance of landing in a resume pool rather than against a named opening.
- **Sibling reqs are dead — do not chase them.** Web results surfaced "Research Platform – Data Platform Engineer" (a closer title match) and "Research Platform Infrastructure Engineer" via Built In. Checked 2026-08-25: the first returns **HTTP 404**, the second is Cloudflare-blocked and absent from the live board. Those Built In listings are stale. **Platform Engineer is the live req.**
## Application Decision
- Audience profile: **International Tech, English** (config default). US market maker, English working language, ex-Google Zurich office. No Swiss/DACH conventions, no photo.
- Evidence Fit: **71/100** (revised down from 74 on 2026-08-26 — see Working Model below)
- Fit class: **Adjacent (mid-band)**
- Hard gate: **QUALIFICATIONS PASS, but a PRACTICAL CONSTRAINT IS UNRESOLVED** — `critique_framework.md` flags a NO-GO when "location… or mandatory relocation is unresolved." The Zurich working model is unknown and the evidence points to 5-day onsite. **Resolve before Phase 2.**
- Channel Strength: **Weak (cold)** — no named contact on the posting, no known connection. Would be the 5th consecutive cold-channel application.
- Cohort slot: Adjacent would be **2/2** (Aker BP holds 1/2). Note: a strict reading of `application_strategy.md` would class this **Stretch** (0/1 free) — see Gap Assessment.
- Decision: **PENDING USER** — Phase 0 only, per user instruction. Phase 1 not started.
## Requirements
### Minimum qualifications — zero Gaps
| # | JD requirement | Class | Evidence |
|---|---|---|---|
| Q1 | Bachelor's in CS or related field | **Direct** | B.Eng. Information & Telecommunication Technologies + M.Eng. Computer Aided Engineering (Software Design & Engineering), UniBw München |
| Q2 | 3+ years professional software engineering | **Direct** | ~11 years (2015–present) |
| Q3 | Strong programming in Python, **or** Go, **or** C++, or similar systems-oriented languages | **Direct** | Python `production-current`. Disjunctive requirement — Python alone satisfies it. C++ is `production-historical-limited` / `allowed-with-context`; usable as secondary colour only, never headlined |
| Q4 | Building **and operating** distributed systems in production | **Direct** | Kafka, Kubernetes, Spark, Airflow carrying production traffic at Swisscom; Component Owner with data quality, incident handling and on-call (SW-1) |
| Q5 | Designing scalable backend services, APIs, and data-intensive applications | **Adjacent** | Data-intensive applications Direct (BS-2 data services over Oracle/Hadoop; SW-4 Python data apps on K8s). Backend services/APIs partial. **SDKs: no evidence** |
| Q6 | Strong understanding of distributed systems, data-intensive applications, system design fundamentals | **Adjacent** | iSAQB CPSA-F (software architecture) + AWS SA-Associate + production operation. But "distributed systems" as a *design discipline* has no canonical entry — 0 hits in `claims.json` |
| Q7 | Cloud platforms (AWS, GCP, Azure) and modern infrastructure | **Direct** | AWS `production-current`. Disjunctive — AWS alone satisfies. (GCP and Azure are `forbidden` in `claims.json`; never list them) |
### Preferred ("is a plus") — complete hit
| JD | Class | Evidence |
|---|---|---|
| Kafka | **Direct** | `production-current` |
| Kubernetes | **Direct** | `production-current-and-historical` |
| Spark | **Direct** | PySpark `production-current` |
| Airflow | **Direct** | `production-current` |
| Distributed databases | **Direct** | Teradata, Redshift, Athena/Iceberg |
**All five preferred technologies are Direct.** This is the strongest raw stack alignment in the application log.
### Core responsibilities
| # | JD responsibility | Class | Note |
|---|---|---|---|
| R1 | **Design and build distributed platforms and services** supporting quantitative research, simulations, AI-powered applications | **Adjacent** | Title-defining. He builds governed data products and pipelines *on* Swisscom's platform; authoring research-platform infrastructure at firm scale is a genuine stretch. **This is the risk.** |
| R2 | Develop scalable platform services and **data systems for ingestion, transformation, storage and lifecycle management of large datasets** | **Direct** | Also title-defining, and a plain description of SW-1 / SW-7 |
| R3 | Backend services, APIs and **SDKs** for researcher productivity and self-service | **Adjacent** | Data products for downstream self-service is adjacent; SDKs are a Gap |
| R4 | Improve **performance, reliability, scalability** of critical research systems in **high-throughput** environments | **Adjacent** | Query performance tuning (BS-2), SLOs (BS-3), on-call reliability (SW-1). "High-throughput" in the HFT sense is not evidenced |
| R5 | Collaborate with researchers to translate requirements into production-ready systems | **Direct** | SW-8 stakeholder requirements; Bosch delivery to analysis teams |
## ATS Keywords
- **Languages:** Python, C++ (context only), SQL
- **Distributed/infra:** distributed systems, Kubernetes, Kafka, Spark, Airflow, containerization, CI/CD
- **Data:** data platform, ingestion, transformation, storage, lifecycle management, large datasets, data-intensive applications, distributed databases, ETL/ELT, data modelling, query performance
- **Cloud:** AWS, S3, Glue, Athena, Iceberg, Redshift, Lambda, Step Functions, CloudFormation
- **Systems:** scalability, reliability, performance, backend services, APIs, production operation, on-call
- **Domain:** quantitative research, simulations, backtesting, analytics workflows, model development
- **Deliberately absent (no evidence):** Go, SDK authoring, low-latency/HFT, GCP, Azure, Terraform
## Gap Assessment
**Four real gaps, in severity order:**
1. **Platform authoring at firm scale (R1) — the decisive one.** The JD wants someone to *design and build the platform researchers use*. Dennis owns components, pipelines and data products *within* Swisscom's company-wide Data Mesh. `CLAUDE.md` Scope Discipline forbids pairing a full-ownership verb with an org-scale object, so this gap **cannot be written around** — it can only be honestly bridged. Same shape as the Aker BP framework-authorship stretch (which scored 79 and still went out as Adjacent).
2. **No professional quantitative/finance domain.** Now partially mitigated by **PP-1** (self-built momentum/signal platform) and **PP-2** (Udacity AI for Trading nanodegree), added to `claims.json` 2026-08-25. These are genuine self-directed depth — "residual 12-1 momentum, ATR stop/trail, quintile gating" is not JD-scraped vocabulary. But PP-1 is single-user and self-hosted: **never call it distributed, production or multi-user, and never quote a PnL, Sharpe, return or backtest figure.**
3. **SDKs — no evidence at all.** 0 hits in `claims.json`. Appears in R3 only, not in the minimum quals, so it is not a gate.
4. **Low-latency / high-throughput systems.** 0 canonical hits. The only throughput reference is the M.Eng. thesis, which `claims.json` `thesis_limits` explicitly flags as a methods prototype with no operational data — **not usable as high-throughput evidence.**
**Non-gaps — do not score these as gaps:**
- **Go.** Q3 is disjunctive ("Python, Go, **or** C++"). Python satisfies it. The "systems-oriented languages" phrasing is a cultural signal about the room, not a requirement.
- **Distributed systems experience (Q4).** A production Kafka/K8s/Spark/Airflow estate *is* a distributed system. Do not undersell this.
**Hard-gate verdict: PASS, with a flagged risk.** No minimum qualification is a Gap. This is materially cleaner than the AWS FDE NO-GO, where the title-defining capability was a flat Gap that `claims.json` globally forbids claiming. Here the title-defining function **splits**: R2 Direct, R1 Adjacent.
**Classification caveat, stated plainly:** `application_strategy.md` says *"Stretch: below 60 **or a defining responsibility is only Adjacent**."* R1 is a defining responsibility and is Adjacent, so a strict reading classes this **Stretch**, not Adjacent. It is recorded as Adjacent because R2 — equally title-defining — is Direct, and because the Aker BP precedent (identical authoring-vs-building shape) was recorded as Adjacent at 79. **Both cohort slots are free either way**, so the label does not block the decision. Flagging it so the choice is deliberate, not accidental.
## Working Model — UNRESOLVED, decision-blocking (raised by user 2026-08-26)
**The JD is silent.** Grep of the verbatim posting for hybrid / remote / onsite / in-office /
days per week / office returns **zero matches**. Phase 0 scored practical constraints 5/5 by
assuming Zurich was workable; that assumption was unexamined and is now withdrawn.
**Evidence, and its quality:**
| Source | Says | Reliability |
|---|---|---|
| Ken Griffin, on record (Fortune 2023, repeated since) | Brought Citadel back **5 days a week**, early and against the grain; calls it his most important leadership decision; all employees in office full-time; remote work "hinders innovation" and costs early-career development | **High** — founder/CEO, on record, repeated, framed as core culture not policy |
| Fishbowl post, Nov 2022 | "usually 3 days minimum… depends on your manager" | **Low** — anonymous, 4 years stale, Operations not Engineering, contradicted by the CEO |
| Citadel Securities Zurich official policy | Not found | — |
**Consequence if 5-day onsite is correct:** Bern → Zurich is ~1h each way, so ~2–2.5h commuting
per day, five days a week — or relocation to Zurich. That breaks the standing constraint
(`user_comp_bar`: Bern-based, hybrid max 2–3 days) and `user_international_mobility` ("keep a
Swiss home base", no relocation).
**Scoring impact:** Practical constraints **5/5 → 2/5** (language and authorization remain clean:
English working language, EU citizen + Swiss B permit). Evidence Fit **74 → 71**. Still Adjacent,
now mid-band rather than one point below Core.
**This must be ASKED, not researched** — same class of question as the SBB Anforderungsniveau K
band. No amount of further searching produces an authoritative answer for the Zurich office
specifically, and the cost of guessing wrong is a full package plus a process built on a
commute the user has already declined elsewhere.
**Recommended sequence: resolve this BEFORE Phase 2, not after.** The SBB package was submitted
with its equivalent question still open, and that question is now a live screening topic instead
of a settled fact. Repeating that pattern here would be a deliberate choice, not an oversight.
## Company Context
- **Citadel Securities** — technology-driven global market maker (equities, options, fixed income). Distinct from Citadel LLC, the hedge fund. Institutional and retail liquidity provision.
- **Zurich office:** Klausstrasse 4, 8008 Zurich. Founded 2019. **Heavily ex-Google.** Platform research is led by **Costas Bekas**, former IBM distinguished researcher. **Ferenc Tóth joined June 2026** as a software engineer on the platform research team, from seven years as a Google staff engineer. Principal research is headed by **Nicolai Meinshausen**, the ETH statistician (returned August 2025).
- **What the Zurich office actually is:** of 10 open Zurich roles, the large majority are quant researchers, ML researchers and quant interns — mostly PhD-gated. Platform Engineer is one of only ~3 experienced non-quant engineering seats there. It is a research-first office, and platform engineering exists to serve researchers.
- **Role purpose (JD's own words):** systems that "power large-scale simulations, analytics, and model development" and let researchers "transform data into actionable insights and alpha-generating trading strategies." Platform work here is directly revenue-attached, which is unusual and worth understanding before an interview.
- **Why them angle:** research-serving data platform work, at a firm where the platform's quality visibly determines research throughput — a sharper version of what he does at Swisscom, in a domain he already follows privately.
## Framing Strategy
- **Lead narrative:** *A production data-platform engineer who builds and operates the ingestion, transformation and storage systems analytical consumers depend on — and who independently built an end-to-end equity signal platform because the domain genuinely interests him.*
- **Reframing map:**
- SW-1 / SW-7 → R2 (ingestion/transformation/storage/lifecycle at scale). **Strongest single mapping; lead with it.**
- SW-4 (Python apps on K8s + GitLab CI/CD) + Kafka/Spark/Airflow → Q4 distributed systems in production.
- BS-2 (data services over Oracle/Hadoop, query performance tuning) → Q5 data-intensive applications + R4 performance.
- BS-3 (Application Owner, SLOs, vendor coordination, training) → R4 reliability + R5 collaboration.
- PP-1 + PP-2 → domain credibility for R1's *"quantitative research, simulations"* and R3's self-service surfaces.
- **Emphasize:** production ownership and on-call (they run revenue systems); the exact Kafka/K8s/Spark/Airflow/AWS stack; self-service data products for downstream consumers; genuine domain interest via PP-1.
- **Downplay:** Data Mesh governance vocabulary (enterprise-flavoured, not their register); Spotfire/BI; ITIL; anything reading as enterprise-IT process.
- **Never claim:** authoring/owning a company-wide platform; distributed-systems *design* at firm scale; SDK authoring; low-latency/HFT experience; professional quant work; any PP-1 performance number.
- **Level note:** JD states "Bachelor's, 3+ years" — he is *above* the stated bar at Staff + Component Owner. Do not down-level the framing to match the JD's floor; the $175k–350k base band is wide enough to span several levels.
## Critique Context
- **Reviewer persona:** a platform engineer or eng manager on Research Platform Engineering — likely ex-Google/ex-big-tech, systems-oriented, in a research-first office. Reads for distributed-systems depth and production instinct, not for enterprise process maturity.
- **Competitive landscape:** brutal. Citadel Securities' engineering bar is among the highest in the market, and the Zurich office is visibly staffed from Google and academia. Realistic competition is ex-FAANG infrastructure engineers and systems specialists. Dennis's differentiator is not raw systems pedigree — it is production data-platform ownership plus real domain interest.
- **Domain vocabulary to get right:** alpha, signal, backtest, simulation, cross-sectional momentum, factor, research throughput, market making. Using these correctly is credibility; using them loosely is worse than not using them.
- **Likely probes:** (1) "Describe a distributed system you designed" — the weakest ground, prepare an honest scoped answer about pipeline architecture rather than claiming platform authorship; (2) "What's your throughput/latency experience?" — genuinely thin, do not bluff; (3) PP-1 — expect real interest and real scrutiny, and expect to be asked about results, which must be answered honestly as "not a validated track record".
- **Compensation caution:** the **$175,000–350,000 base range in the posting is a US pay-transparency disclosure** (NY law) and says nothing about the Zurich figure. Do not plan against it. Citadel Securities Zurich very likely clears the CHF 180k all-in bar, but that is inference, not evidence — ask.
## Cover Letter Decision
- **Decision: YES (recommended), if Phase 1 proceeds.**
- Rationale: this is the rare case where a letter carries information the resume structurally cannot. The resume can list PP-1 only as a compact project line; it cannot convey *why* a Swisscom data engineer credibly wants a research-platform seat at a market maker. The domain-interest story is the single strongest differentiator against an ex-FAANG competitive field, and it needs prose.
- Institution type: US market maker, research-first office. Register: direct, technical, unsentimental. No enterprise formality, no Swiss/DACH conventions.
- Structure: (1) what he builds now, mapped to R2 in their words; (2) production ownership and on-call as the reason he thinks about reliability the way a revenue-system team does; (3) PP-1/PP-2 as honest self-directed domain grounding — explicitly labelled a personal project, no numbers; (4) short close on Zurich, no relocation needed, EU citizen + Swiss B permit.
- Length: 1 page, 4 short paragraphs.
- **Hard constraints:** never imply professional quant experience; never quote a PP-1 result; never claim platform authorship; check `bundle_data_engineer.md` §S5 before reuse — its opening hook contained an SW-1 scope violation (fixed 2026-08-21, but verify).
## Resume Plan (Phase 1 — proposed 2026-08-26, awaiting confirmation)
Bundle: **`bundle_data_platform.md`** primary (Tier 3), `bundle_data_engineer.md` secondary.
Format: 2-page resume, International Tech, English.
Bullet lengths: **natural** per `resume_reference.md` §7 — the SKILL.md "all bullets 2L" line is stale,
superseded by the 2026-07-27 global correction. §7a cadence variety applies.
### Swisscom — Staff Data, Analytics & AI Engineer (5 bullets)
| | ID | Achievement | JD match |
|---|---|---|---|
| * | SW-7 | Governed data products in the company-wide Data Mesh; onboard sources; metadata + lineage in Atlassian Compass | **R2 Direct** — lifecycle mgmt; R3 self-service |
| * | SW-1 | Migrated owned Fulfillment/Product Analysis pipelines onto AWS (Glue, Athena/Iceberg, Redshift, Airflow, CloudFormation) | **R2 Direct**, Q7 cloud |
| * | SW-2 | Component Owner, business-critical Fulfillment ETL (Oracle/Kafka → Teradata): data quality, governance, incidents, on-call | **Q4 Direct**, R4 reliability |
| * | SW-3 | Build and operate Python data applications on Kubernetes with GitLab CI/CD | **Q4 Direct**, Q5 backend services |
| o | SW-4 | Deliver data products and analyses with internal stakeholders; translate requirements | R5 Bridge |
### Bosch Semiconductor Dresden — Senior Engineer, Data Analysis (4 bullets)
| | ID | Achievement | JD match |
|---|---|---|---|
| * | BS-2 | Data services in Python/Java/C# over Oracle and Hadoop/Impala; **query performance tuning**; fab sensor/process data | **Q5 Direct**, R4 performance |
| * | BS-1 | Integrated containerized ML inference (Docker/K8s/Ansible) into a 24/7 fab | **Q4 Direct**, R1 "AI-powered applications" |
| * | BS-4 | ELK/Kafka anomaly-detection PoC with Grafana/Prometheus/Loki monitoring | Kafka Direct, R4 reliability |
| o | BS-3 | Application Owner: SLOs, vendor coordination, training and documentation | R4 + R5 Bridge |
### Vizrt — Test Automation / DevOps Engineer (1 bullet) — **user decision needed**
| | ID | Achievement | JD match |
|---|---|---|---|
| ? | VZ-1 | Contributed **Python and C++** engineering to a **distributed video-transcoding backend** | **Q3 C++ bonus + Q4/Q6 distributed** |
| x | VZ-2 | Automated A/V integration tests as CI/CD quality gates | Weak for this JD |
**VZ-1 is unusually well-matched to this specific JD** — it is the only canonical evidence that pairs
C++ with a genuinely distributed backend, which is exactly Q3's bonus clause plus Q6. **But the user
explicitly skipped the Vizrt low-latency angle on the Snowflake package.** Raising it rather than
assuming. Verb must stay hedged ("contributed"); broadcaster names are `forbidden`.
### Fraunhofer CML — Research Software Engineer (1 bullet)
| | ID | Achievement | JD match |
|---|---|---|---|
| * | FC-3 | Developed containerized microservices for the MISSION **research** data-exchange platform | R1/R3 Bridge — closest canonical analogue to "research platform" |
| o | FC-1 | Jenkins CI/CD setup | Weaker here than FC-3 |
### Generali (0–1 bullets)
| | ID | Achievement | JD match |
|---|---|---|---|
| x | GN-1 | BDD test automation PoC, Jenkins ownership, training | Weak — recommend omit |
### Selected Project — PP-1 (1–2 bullets) — **new section, structural decision needed**
| | ID | Achievement | JD match |
|---|---|---|---|
| * | PP-1 | Self-built equity signal platform: price/fundamental/sentiment ingestion, cross-sectional momentum book, scheduled scan + backtest pipelines, dashboard + alerts | **Domain credibility for R1 "quantitative research, simulations"** |
| o | PP-2 | Udacity *AI for Trading* nanodegree | Certification-context only |
The current 2-page template has **no Projects section**. Adding one is a Phase 2 structural change.
PP-1 is the main differentiator against an ex-FAANG field, so it should appear — but it must be
labelled a personal project, and `claims.json` forbids calling it distributed/production/multi-user
or quoting any PnL, Sharpe, return or backtest figure.
### Budget
| Set | Bullets |
|---|---|
| Recommended (`*`) | **11** |
| With `o` options (SW-4, BS-3, PP-2) | 14 |
| Target (`resume_reference.md` §4) | 11–14 |
**Budget: PASS** at both ends.
### Forced exclusions (provenance)
- **SW-5 Security Champion** — omit by default; JD has no security requirement.
- **SW-8 LLM work** — configuration-level only; `forbidden` to imply built/deployed LLM systems. LLM
sentiment in PP-1 is fine as personal-project scope.
- **GCP / Azure / Terraform / TypeScript / FastAPI** — `forbidden` in `claims.json`; never list even
though the JD names GCP and Azure (Q7 is disjunctive; AWS satisfies it).
- **M.Eng. thesis throughput figure** — barred by `thesis_limits`; not usable as high-throughput evidence.
## Output Files
- `e2e_citadel_securities_platform_engineer_resume.tex` / `.pdf` — 2 pages, 13 bullets
- `e2e_citadel_securities_platform_engineer_cover_letter.tex` / `.pdf` — 1 page, 270 words, 3 paragraphs
- `resume.cls` (copied from templates)
- `JD_CitadelSecurities_PlatformEngineer.txt` — verbatim posting
**Phase 2 decisions taken (user said "lets build the resume" without answering the two open items):**
- **VZ-1 included.** Only canonical evidence pairing C++ with a distributed backend; this JD names
both explicitly. Hedged verb "Contributed" per `claims.json`; no broadcaster names. Reversible.
- **Personal Project section added** for PP-1, labelled three ways (heading, right-aligned
"Personal project", and "not a production or multi-user service" in the body). No performance
figure of any kind. PP-2 appears as "Supporting coursework", never as a credential.
- **Generali kept** with one bullet, against the Phase 1 "omit" recommendation — omitting it left an
unexplained May 2015–Jun 2017 employment gap, which costs more than the weak bullet does.
**Verification:**
- `validate_resume_system.py --document` → **PASS**, 1 warning (cadence, see below)
- Compiled with MiKTeX pdflatex, **exactly 2 pages**, rendered PDF inspected: no clipping, orphans,
header wrapping or bad breaks; umlauts, `C\#` and `C++` all render correctly
- Page 2 fills well; page 1 has bottom whitespace from the deliberate `\newpage`. Not padded —
`resume_reference.md` §9: "Do not add content merely to reduce bottom whitespace."
**Known residual — cadence 7/12 (58%), above the §7a ~50% guide.** Reduced from 75% by reshaping two
bullets. The remaining flags are almost all legitimate technology enumerations ("Oracle, Kafka,
Python and Teradata"; "Elasticsearch, Logstash, Kibana and Kafka"). **The §7a checker cannot
distinguish a rhetorical rule-of-three from a list of tools actually used** — a real limitation of
the pattern. Cutting these would mean deleting accurate, ATS-relevant technology names, which §7a
itself forbids ("never trade accuracy… to fix cadence"). Left as-is deliberately.
**Trap hit and fixed during generation:** the first draft named the non-canonical tools in a header
comment explaining why they were excluded. The validator scans **raw file text**, so it errored on
all three. This is the exact failure the SBB `.tex` header warns about. Comment rewritten without
naming them.
## Edit History
### Edit 1 (2026-08-26): Tier 1 relevance fixes
- **Source:** `critique_citadel_securities_platform_engineer.md` Tier 1 items 1-3. User approved
"tier 1 fixes" only; the Tier 2 headline retitle was deliberately NOT applied.
- **Changes (all MODIFY, no budget change):**
1. SW-7 bullet -> now carries **ingestion** and **lifecycle**: "from source-system ingestion
through to the metadata and lineage in Atlassian Compass that keep each product discoverable
across its lifecycle."
2. SW-1 bullet -> now carries **transformation** and **storage**: "transformation and storage on
Glue, Athena with Apache Iceberg and Redshift, orchestration in Airflow". Scope hedge
("contributing to the wider company migration programme") retained.
3. SW-2 bullet -> names the estate a **distributed system** he operates, not designed.
4. Skills line 2 relabelled **"Distributed systems"** and now names **distributed databases
(Teradata, Redshift, Athena/Iceberg)** - closes the fifth preferred-technology hit that Phase 0
claimed as Direct but the document never wrote.
- **Not added, deliberately:** scalable (unverified scale claim), SDK / high-throughput /
data-intensive / simulation / model development / self-service (zero canonical evidence).
Verified still absent after the edit.
- **Verification:** validator **PASS, 0 warnings** (the cadence warning cleared as a side effect -
restructuring SW-1/SW-2 with em-dashes broke the comma-list pattern). Compiles to exactly
**2 pages**, zero overfull/underfull boxes, PDF re-inspected clean.
| Metric | Before | After | Delta |
|---|---|---|---|
| Pages | 2 | 2 | 0 |
| Validator errors | 0 | 0 | 0 |
| Validator warnings | 1 (cadence 58%) | **0** | -1 |
| Bullets | 13 | 13 | 0 |
| JD term coverage | 19/37 (51%) | **25/37 (68%)** | +6 |
**Estimated Document Quality ~91/100** (relevance 10 -> ~14, mechanics 9 -> 10). Estimate only -
not a re-scored critique.
## Status
- Phase 0: **DONE** (2026-08-25)
- Phase 1: **DONE** (2026-08-26) — 13 bullets confirmed
- Phase 2: **DONE** (2026-08-26) — resume compiled and verified
- Cover Letter: **DONE** (2026-08-26) — 1 page, 270 words, 3 paragraphs, validator PASS.
No external hooks by design (`cl_reference.md`: omit an unnecessary hook); the only hook is the
JD's own first-party language. States the R1 gap plainly rather than hiding it. PP-1 labelled
personal with an explicit "I make no claims for its results".
- Critique: **STALE by design** — scored 86/100 on the pre-Edit-1 resume and before the cover
letter existed. User elected to submit as-is rather than re-critique (2026-08-26). Estimated
post-edit Document Quality ~91; not re-scored, and that estimate must not be quoted as a score. — **Document Quality 86/100**, Evidence Fit 71, Channel Weak.
No Tier 1 truth findings; validator PASS. Three Tier 1 relevance fixes (~+4), all honestly
available: (1) R2's own vocabulary — ingestion/transformation/storage/lifecycle — is absent
despite R2 being the single Direct core responsibility; (2) "distributed systems" never appears
as a phrase though Q4 is a required qual scored Direct; (3) "distributed databases" claimed as a
Direct preferred hit but never written. File: `critique_citadel_securities_platform_engineer.md`
- **STILL BLOCKED on the working-model question** — resume built at user direction with that
question open. If Zurich is 5-day onsite, this package does not get submitted.
## SUBMITTED 2026-08-26 — and a finding that outranks the working-model question
**Application sent.** Cohort entry and decision log both recorded.
### The form had no Zurich option
The user reports that the application form's **preferred-work-location selector offered no Zurich**,
despite:
- the JD body reading "in Miami, **Zurich** or New York";
- the posting header listing "New York, Miami, **Zurich**";
- the site's own **Zurich location filter returning this exact req** (verified 2026-08-25, 10 Zurich roles).
The careers site footer reads **"Citadel Enterprise Americas LLC"**, so the most likely explanation is
that the apply flow is a US-entity form and European seats route elsewhere — or that the Zurich seat
is closed while the multi-location posting has not been updated. A form bug is possible but least likely.
**Consequence.** This partly supersedes the working-model risk. If there is no Zurich seat reachable
through this form, whether Zurich is hybrid or 5-day onsite is moot — the application may be sitting
against a US requisition, which is a hard NO on relocation (`user_international_mobility`: no
relocation, keep a Swiss home base).
**Carry-forward rule for any future Citadel Securities application: check the application form's
location selector BEFORE investing in a package.** The posting's stated locations and the site's own
filter both proved unreliable as evidence that a Zurich seat is actually reachable.
### Expectation
User submitted with low effort on the form fields and low expectation of a reply. That is a
reasonable read: cold channel, an evergreen standing req, an ex-Google/ETH-staffed office, and now a
location mismatch in the apply flow itself.
### Package status
Documents are finished and reusable. Nothing about the package is the weak point here — the resume
and letter both cleared the validator with zero warnings and the Tier 1 fixes lifted JD coverage
51% → 68%. If a better-matched Citadel or market-maker req appears, the package retargets cheaply.
## Finalization (2026-08-26)
Per `shared_ops.md` → Finalization:
1. ✅ Session, resume `.tex`/`.pdf`, critique all present
2. ✅ Cover letter `.tex`/`.pdf` present (session decision YES)
3. ✅ Canonical validator re-run on **both** submitted documents → **PASS, 0 warnings each**
4. ✅ Final PDFs copied to `Dennis_Thiessen_Resume.pdf` and `Dennis_Thiessen_Cover_Letter.pdf`
5. ✅ **Submission recorded 2026-08-26** in `job_scout/state/decisions.json`
6. ✅ **Cohort entry added** — Adjacent **2/2** (cohort now 4/10; channel weak ×3)
**On confirmation of submission, run both:**
```
./job_scout/.venv/Scripts/python.exe resume_builder/helpers/cohort_tracker.py add --company "Citadel Securities" --role "Platform Engineer, Research Platform Engineering (Zurich)" --date <YYYY-MM-DD> --fit-class adjacent --evidence-fit 71 --channel weak --hard-gate pass --outcome applied
./job_scout/.venv/Scripts/python.exe job_scout/scout.py --decide "https://www.citadelsecurities.com/careers/details/platform-engineer/" applied "<note>"
```
**Carry-forward risk — unchanged by submission.** The Zurich working model is still unknown and the
evidence points to 5-day onsite. **Ask it in the first recruiter contact.** If it is 5 days, this is
a decline on the Bern constraint regardless of how the process goes.
- Next: ask the working-model question at first contact; record submission when sent
- Resume: PENDING
- Cover Letter: PENDING (decision YES)
- Critique: PENDING
- Next: user decision on whether to proceed
output/KdoCy_DevOps_Data_Platform/JD_kdocy_devops_data_platform.txt
+130
View File
@@ -0,0 +1,130 @@
https://jobs.admin.ch/offene-stellen/devops-engineer-iii-data-platform/752ded37-1753-4ac8-946a-b0c958d4424a
DevOps Engineer III (Data Platform) - Schweizer Armee - Kommando Cyber Kdo Cy
-->
Übersicht
Kontakt
Über uns
Benefits
Ähnliche Stellen
Stellen teilen
Drucken
Job-Abo einrichten
Später bewerben
Interessiert?
Möchten Sie sich bewerben? Dann senden Sie sich die Stelle per E-Mail zu und bewerben Sie sich bequem von zu Hause aus.
E-Mail
Link senden
Der Link zu dieser Stelle wurde an die eingegebene E-Mail-Adresse verschickt.
DevOps Engineer III (Data Platform)
Die Schweiz mitgestalten
DevOps Engineer III (Data Platform)
Schweizer Armee - Kommando Cyber Kdo Cy
Zimmerwald | 80-100% (Jobsharing möglich)
Eintrittsdatum:
nach Vereinbarung
Anstellungsart:
unbefristet
Referenz-Nr.:
JRQ$540-19848
Arbeitsort:
Eichacher, 3086 Wald
Diesen Beitrag kannst du leisten
Moderne IaC- und GitOps-Lösungen planen, implementieren und betreiben sowie den gesamten Deployment-Zyklus verantworten
MLOps-Umgebungen für das Deployment, Monitoring und den zuverlässigen Betrieb von KI- und Machine-Learning-Modellen entlang ihres gesamten Lebenszyklus bis hin zur GPGPU-Hardware entwerfen, betreiben, automatisieren und weiterentwickeln
Komplexe Linux-Systemlandschaften für den Ende-zu-Ende-Betrieb einer IKT-Schicht für eine Datenplattform administrieren, warten und kontinuierlich optimieren, um einen stabilen und performanten Betrieb sicherzustellen
Linux-Systeme ganzheitlich administrieren, Performance optimieren sowie System-Hardening und Security-Konfigurationen umsetzen
Monitoring-, Logging- und Observability-Lösungen betreiben, weiterentwickeln und optimieren, um einen zuverlässigen Systembetrieb sicherzustellen
Die physische Infrastruktur im Rechenzentrum mit Server-, Netzwerk- und Datacenter-Komponenten betreuen sowie Installationen und Wartungsarbeiten unterstützen
Das macht dich einzigartig
Abgeschlossene Ausbildung im IT-Bereich HF/FH sowie mehrere Jahre Praxis im Engineering und im Betrieb von komplexen IT Systemen
Fundierte DevOps-Kompetenz durch mehrjährige Erfahrung im Aufbau, Betrieb und in der Automatisierung moderner Linux-Infrastrukturen sowie umfassende Kenntnisse in Infrastructure-as-Code, GitOps und zugehörigen Prozessen
Streben nach Exzellenz mit hoher Lernbereitschaft und Offenheit für ganzheitliche, elegante IKT-Lösungen sowie innovative Ansätze abseits des Mainstreams wie Dhall, Pull-GitOps und «Dew-» sowie «Fog-Computing»-Paradigmen
Hands-on-Mentalität sowie eine selbstständige und teamorientierte Persönlichkeit mit Freude an der Arbeit über den gesamten Technologie-Stack hinweg – von der physischen Infrastruktur im Rechenzentrum bis hin zu automatisierten Plattform- und Big-Data-Umgebungen wie ClickHouse und Kafka
Ein tadelloser Leumund sowie die Fähigkeit, sich in ein nachrichtendienstliches Umfeld zu integrieren und dabei jederzeit diskret zu handeln, werden vorausgesetzt
Aktive Kenntnisse einer zweiten Amtssprache sowie gute Englischkenntnisse
Auf den Punkt gebracht
Du bist Teil eines Data-Science-Teams im Bereich der Verarbeitung von erfassten zivilen und militärischen Kommunikationsdaten. Die daraus gewonnenen nachrichtendienstlichen Erkenntnisse unterstützen den NDB und den MND und leisten damit einen wichtigen Beitrag zur Sicherheitspolitik der Schweiz. Als DevOps Engineer III bist du fachlich mitverantwortlich für den Entwurf, die Umsetzung und den Betrieb einer grossen, verteilten Dateninfrastruktur.
Eintrittsdatum:
nach Vereinbarung
Anstellungsart:
unbefristet
Referenz-Nr.:
JRQ$540-19848
Arbeitsort:
Eichacher, 3086 Wald
Fragen zur Stelle
Marcel Matthey-Doret
Chef Centers of Excellence CEA
+41 58 46 78409
Jetzt bewerben
Sind Sie aktuell beim Bund angestellt?
Ja
Nein
Fragen zur Stelle
Marcel Matthey-Doret
Chef Centers of Excellence CEA
+41 58 46 78409
Auf den Punkt gebracht
Du bist Teil eines Data-Science-Teams im Bereich der Verarbeitung von erfassten zivilen und militärischen Kommunikationsdaten. Die daraus gewonnenen nachrichtendienstlichen Erkenntnisse unterstützen den NDB und den MND und leisten damit einen wichtigen Beitrag zur Sicherheitspolitik der Schweiz. Als DevOps Engineer III bist du fachlich mitverantwortlich für den Entwurf, die Umsetzung und den Betrieb einer grossen, verteilten Dateninfrastruktur.
Dein Einsatz für Sicherheit und Freiheit
Das Kommando Cyber (Kdo Cy) steht für Kompetenz in den Bereichen Cyber und Digitalisierung. Wir vereinen die Begeisterung für unsere Aufgabe mit Know-how, Innovation, modernsten Technologien und agilen Arbeitsweisen. Wir erbringen sichere, robuste und resiliente Informatikleistungen für die Schweizer Armee und unsere Partner im Sicherheitsverbund Schweiz. Damit ermöglicht das Kdo Cy der Schweizer Armee den notwendigen Wissens- und Entscheidvorsprung – für eine sichere Schweiz.
Die Schweizer Armee beschäftigt über 9'000 Mitarbeitende in vier militärischen und rund 200 zivilen Berufen – und dies an 110 Standorten schweizweit. Auch für den Nachwuchs ist gesorgt: Die Schweizer Armee bildet in über 30 Berufen Lernende aus.
Weiter
Zurück
Link zu mehr Informationen
Das bieten wir
#F7B4B8
Arbeiten für die Schweiz
Wir setzen uns für das Erfolgsmodell Schweiz ein und arbeiten zum Wohl der Bevölkerung.
#F7B4B8
Arbeiten für die Schweiz
Wir setzen uns für das Erfolgsmodell Schweiz ein und arbeiten zum Wohl der Bevölkerung.
#B8E1D8
Gelebte Vielfalt
Dank Chancengleichheit entfalten wir unsere Kompetenzen und bringen unterschiedliche Perspektiven ein.
#B8E1D8
Gelebte Vielfalt
Dank Chancengleichheit entfalten wir unsere Kompetenzen und bringen unterschiedliche Perspektiven ein.
#FFF0BB
Gesund am Arbeitsplatz
Wir unterstützen und beraten unsere Mitarbeitenden im Bereich der physischen und psychischen Gesundheit.
#FFF0BB
Gesund am Arbeitsplatz
Wir unterstützen und beraten unsere Mitarbeitenden im Bereich der physischen und psychischen Gesundheit.
Alle Benefits
Fragen zur Stelle
Marcel Matthey-Doret
Chef Centers of Excellence CEA
+41 58 46 78409
Jetzt bewerben
Jetzt bewerben
Sind Sie aktuell beim Bund angestellt?
Ja
Nein
-->
Zusätzliche Informationen
Eine Anstellung bei der Schweizer Armee macht Sinn und bringt viele Vorteile mit sich. Überzeuge dich auf 
armee.ch/vorteile
.
Interessiert an Jobsharing? Du kannst dich einzeln oder mit Partner/-in bewerben. Auch wenn ihr euch im Duo bewerbt: Reicht eure Bewerbungen einzeln ein und gebt den Namen der Partnerin/des Partners im Bewerbungskonto an.
Angebote von Personalvermittlern werden nicht berücksichtigt.
Die nächsten Schritte
Frauen sind in unserer Verwaltungseinheit noch untervertreten. Ihre Bewerbungen sind daher besonders willkommen.
Ähnliche Stellen
Staatssekretariat für Wirtschaft SECO
ICT DevOps Engineer Data & Analytics Senior
Bern, Schweiz | 100-100% (Jobsharing möglich)
Bundesamt für Informatik und Telekommunikation BIT
ICT-System-Ingenieur/-in Senior Container Plattform
Zollikofen, Schweiz (und Homeoffice) | 80-100%
Bundesamt für Informatik und Telekommunikation BIT
Datenbank-Administrator/-in Senior (1025)
Zollikofen, Schweiz (und Homeoffice) | 80-100%
Rechtliche Grundlagen
Job-Abo einrichten
Rechtliche Grundlagen
Jetzt bewerben
Sind Sie aktuell beim Bund angestellt?
Ja
Nein
output/KdoCy_DevOps_Data_Platform/critique_kdocy_devops_data_platform.md
+156
View File
@@ -0,0 +1,156 @@
# Critique v3: Kdo Cy, DevOps Engineer III (Data Platform)
**Stand:** 27. August 2026, nach Edit 2 mit Bundeswehr-Offizierslaufbahn und vollständigem Sprachpass.
> **Submission note, 27. August 2026:** Der Nutzer hat die Bewerbung nach Prüfung dieser Critique als gut genug bewertet und unverändert eingereicht. Der Tier-1-Fund zu *„ohne manuellen Eingriff“* bleibt als bekannte, bewusst akzeptierte Abweichung dokumentiert; er wurde nicht nachträglich aus den eingereichten Dokumenten entfernt.
| Achse | Aktuelles Ergebnis |
|---|---|
| Evidence Fit | **79/100, Core (unterer Bereich)** |
| Hard Gate | **PASS** |
| Document Quality | **92/100** |
| Channel Strength | **Weak** |
| Einreichungsstatus | **Noch nicht freigeben: ein Tier-1-Claim muss korrigiert werden** |
Die Offizierslaufbahn ist ein echter Differenzierungsfaktor für diesen Arbeitgeber. Sie verbessert die Glaubwürdigkeit der Motivation und die militärische Anschlussfähigkeit, ersetzt aber keine technische Militär-, Nachrichtendienst- oder Clearance-Evidenz. Deshalb bleibt der Evidence-Fit-Score nach unabhängiger Neubewertung bei 79.
## 1. Fit Verdict und Hard-Gate-Audit
**Verdict: Core, Hard Gate PASS.** Dennis deckt den Titelkern inzwischen ausreichend ab: professionelle Linux-Administration und Ansible-Automatisierung bei Bosch, aktueller Datenplattformbetrieb bei Swisscom, Kubernetes/CI/CD, Kafka, Observability und containerisiertes ML-Deployment. Die einzelnen Belege liegen nicht alle in derselben aktuellen Rolle, ergeben zusammen aber ein belastbares DevOps-/Data-Platform-Profil.
| Bereich | Einstufung | Begründung |
|---|---|---|
| IT-Abschluss und Betrieb komplexer Systeme | Direct | M.Eng.; langjährige Engineering- und Betriebspraxis |
| Linux-Infrastruktur und IaC | Direct-to-Adjacent | Bosch Linux/Ansible sowie CloudFormation sind Direct; das vom JD genannte Pull-GitOps bleibt Gap |
| MLOps und GPGPU | Adjacent | Produktive ML-Inferenzintegration ist stark; GPGPU und vollständiger ML-Lebenszyklus sind nicht belegt |
| Data Platform und Big Data | Direct-to-Adjacent | Kafka, Datenstrecken, AWS, Hadoop/Impala und Data Products sind Direct; ClickHouse fehlt |
| Physische Rechenzentrumsarbeit | Gap | Kein belegter Umgang mit Server-, Netzwerk- oder Datacenter-Hardware |
| Nachrichtendienstliches Umfeld / PSP | Constraint, geklärt | Deutsche Staatsangehörigkeit und B-Bewilligung sind laut Nutzerangabe zulässig; eine PSP wird nicht behauptet |
| Zweite Amtssprache | Soft Gap | Vom Nutzer als Wunschkriterium geklärt; bleibt ein Wettbewerbsvorteil anderer Kandidaten |
| Militärischer Kontext | Adjacent | Sechsjährige Offizierslaufbahn ist Direct als Organisationskontext, aber kein Nachweis für Nachrichtendienst, Cyber, Einsatz oder Clearance |
Der weiterhin offene physische Datacenter-Anteil ist ernst, aber nicht allein titeldefinierend. Die Dokumente behandeln ihn korrekt als Motivation und nicht als vorhandene Berufserfahrung.
## 2. Evidence Fit
| Dimension | Gewicht | Score | Begründung |
|---|---:|---:|---|
| Required qualifications | 35 | **30** | Abschluss, Engineering, Linux/IaC und Englisch sind stark; Pull-GitOps, physische Infrastruktur und zweite Amtssprache bleiben schwächer |
| Core responsibilities | 25 | **19** | Betrieb, Automatisierung, MLOps, Datenplattform und Observability sind belegt; RZ-Hardware, GPGPU und vollständiges GitOps fehlen |
| Level and ownership | 15 | **10** | Component Owner und Application Owner passen; Engineer III wirkt gegenüber aktuellem Staff-Scope lateral bis leicht darunter |
| Recency and depth | 10 | **9** | Kubernetes, Kafka, Python, AWS und CI/CD sind aktuell; die tiefste Linux-/Ansible-Evidenz endet 2022 |
| Domain/tool transfer | 10 | **7** | Impala/Oracle übertragen sich auf ClickHouse-Grundmuster; CI/CD ist kein Pull-GitOps. Die Offizierslaufbahn stärkt den Kontext, nicht die technische Funktionsdeckung |
| Practical constraints | 5 | **4** | Kurzer Arbeitsweg, Arbeitserlaubnis und unbefristete Stelle passen; Lohnklasse und Level bleiben offen |
| **Total** | **100** | **79** | **Core, unterer Bereich; Hard Gate PASS** |
## 3. Document Quality
| Dimension | Gewicht | Score | Begründung |
|---|---:|---:|---|
| Truth and provenance | 25 | **23** | Fast alle Claims sind sauber scoped. Ein absoluter Ergebnisclaim bei Bosch überschreitet die kanonische Evidenz und ist Tier 1 |
| Information hierarchy | 20 | **19** | Zielprofil, Swisscom, Linux/Ansible und Bundeswehr-USP sind sofort sichtbar. Das Kurzprofil belegt allerdings fünf gerenderte Zeilen statt der vorgesehenen zwei bis drei |
| Bullet evidence and impact | 20 | **17** | Technisch konkret und ownership-sicher; kaum verifizierte Kennzahlen. Einige Bullets sind listenlastig, der Bosch-Resultatclaim ist zu absolut |
| Relevance and terminology | 15 | **14.5** | Sehr gute Deckung von Linux, IaC, MLOps, Deployment, Kafka, Observability und Betrieb; echte Gaps werden nicht mit Keywords kaschiert |
| Skills evidence | 10 | **9.5** | Alle sichtbaren Skills sind belegbar und nach Quelle kontextualisiert; C++ und JavaScript wurden sinnvoll aus dem Skills-Block entfernt |
| Mechanics and readability | 10 | **9** | Saubere PDFs, gute Extraktion, keine Box- oder Layoutfehler. Kleine Abzüge für eine unnatürliche IaC-Klammerung, listenreiche Kadenz und zwei aufeinanderfolgende `Entwickelte`-Einstiege |
| **Total** | **100** | **92** | **Hohe Dokumentqualität, aber noch nicht freigabefähig wegen Tier 1** |
## 4. Tier-1-Fund
### Bosch: vollständige Eliminierung manueller Arbeit ist nicht belegt
Aktuell steht:
> „... ermöglichte damit die automatisierte bildbasierte Defektklassifikation **ohne manuellen Eingriff**.“
`BS-1` belegt eine qualitative Reduktion manueller Klassifikation, ausdrücklich aber keinen genauen Umfang. „Ohne manuellen Eingriff“ behauptet eine vollständige Eliminierung und ist damit stärker als die kanonische Evidenz.
**Sichere Korrekturrichtung:**
> „... ermöglichte damit die automatisierte bildbasierte Defektklassifikation und reduzierte den manuellen Klassifikationsaufwand.“
Das ist der einzige aktuelle Tier-1-Punkt. Nach der Korrektur müssen Validator, Kompilierung, Textextraktion und visuelle PDF-Prüfung erneut laufen.
## 5. Competitive Read
Der offensichtliche Konkurrenzkandidat ist ein Senior Linux-/DevOps-Engineer aus Bund, RUAG, armasuisse oder einem Schweizer ISP: aktuelle Linux-Fleet-Verantwortung, Pull-GitOps, Datacenter-Hardware, mögliche PSP-Erfahrung und Deutsch plus Französisch.
Dennis gewinnt dagegen auf der Verbindung aus Data Platform, Kafka, produktivem Pipeline-Ownership, ML-Inferenzintegration, selbst entwickelten Ansible-Erweiterungen und echter militärischer Biografie. Die Offizierslaufbahn macht die Motivation für Kdo Cy deutlich glaubwürdiger als bei einem rein kommerziell geprägten Bewerber. Der Konkurrenzkandidat bleibt dennoch direkter auf den drei schwächsten Achsen: Datacenter, Pull-GitOps und zweite Amtssprache.
## 6. Reader Sequence
| Leser | Verdict | Wahrscheinlicher Eindruck |
|---|---|---|
| Parser / ATS | **Pass** | Arbeitgeber, Titel, Daten und Kernbegriffe werden sauber extrahiert |
| Recruiter | **Forward** | Staff-Scope, lokale Verfügbarkeit, Deutsch, B-Bewilligung und Offizierslaufbahn ergeben eine plausible Kandidatur |
| Hiring Manager | **Interview** | Bosch plus Swisscom decken den technischen Kern ausreichend; die Bundeswehr-Laufbahn stärkt Motivation und Umfeldfit |
| Technical reviewer | **Interview mit Probeachsen** | Erste Fragen werden aktuelle Linux-Tiefe, Pull-GitOps-Abgrenzung, RZ-Hardware und die tatsächliche Wirkung der ML-Automatisierung betreffen |
Die separate HR-Risikoachse bleibt **Borderline**: Lohnklasse, Engineer-III-Level und mögliche Überqualifikation sind nicht durch Textarbeit lösbar.
## 7. Canonical Claim Audit
| Claim im Paket | Evidenz | Einordnung | Sicher? | Massnahme |
|---|---|---|---|---|
| Über zehn Jahre technische Berufserfahrung | Employment history | Direct | Ja | Keine |
| Sechsjährige Offizierslaufbahn, Lehrgang, Offizierschule, Leutnant | BW-1 | Direct | Ja | Keine Führung, Einsätze, Cyberarbeit oder Clearance ergänzen |
| Swisscom Component Owner mit Betrieb, Datenqualität, Incidents und Pikett | SW-2 | Direct | Ja | Keine |
| Python-Datenanwendungen auf Kubernetes mit GitLab CI/CD | SW-3 | Direct | Ja | Keine |
| Migration der eigenen Domänenpipelines auf AWS | SW-1 | Direct, scoped | Ja | Objekt bleibt auf Fulfillment und Product Analysis begrenzt |
| Data Products und Metadaten im unternehmensweiten Data Mesh | SW-7 | Direct, scoped | Ja | Keine Plattform-/Mesh-Ownership behaupten |
| Linux-Administration und Ansible-Automatisierung bei Bosch | BS-6 | Direct | Ja | Keine Datacenter- oder Fleet-Scale-Zahl ergänzen |
| Ansible-Credential-Plugin mit lokaler Zwischenspeicherung | BS-6 | Direct | Ja | Performance/Skalierbarkeit bleibt qualitativ |
| ML-Inferenzintegration in laufender Halbleiterfertigung | BS-1 | Direct | Teilweise | **„ohne manuellen Eingriff“ entfernen** |
| ELK/Kafka-PoC plus Monitoring/Logging/Alarmierung | BS-4 | Direct-to-Adjacent | Ja | PoC-Scope beibehalten |
| Datenservices und Application-Owner-Verantwortung | BS-2, BS-3 | Direct | Ja | Keine |
| Fraunhofer Microservices und Jenkins-CI/CD | FC-3, FC-1 | Direct | Ja | Keine |
| Vizrt Backend-Komponenten und CI/CD-Quality-Gates | VZ-1, VZ-2 | Direct, komponentenspezifisch | Ja | Keine Kundenlieferung oder Gesamtsystem-Ownership ergänzen |
| Generali BDD- und Jenkins-Verantwortung | GN-1 | Direct | Ja | Keine |
| Privater Debian-Server, Härtung und Dienste | PP-3 | Direct, persönlich | Ja | Persönlichen, nicht betrieblichen Kontext beibehalten |
| Motivation: Rückkehr in ein militärisches Umfeld und Beitrag zum Schutz der Schweiz | Nutzerintention plus BW-1 | Motivation, keine Leistungsbehauptung | Ja | Militärischen Kontext sprachlich präzisieren, siehe Tier 2 |
## 8. Tiered Improvements
### Tier 1
1. Bosch-Bullet von der unbelegten vollständigen Automatisierung auf die belegte Reduktion manueller Klassifikation zurückstufen.
### Tier 2
1. **Kurzprofil komprimieren.** Die Sektorenliste ist weniger wertvoll als die bereits sichtbaren Arbeitgeber und macht aus dem Profil fünf gerenderte Zeilen. Ziel: drei bis vier Zeilen bei Erhalt von Swisscom, Bosch und Bundeswehr. Erwarteter Gewinn: **+0.5 DQ**.
2. **IaC-Formulierung glätten.** „CloudFormation als Infrastructure as Code, IaC“ liest sich technisch korrekt, aber sprachlich konstruiert. Besser: „CloudFormation für Infrastructure as Code (IaC)“. Erwarteter Gewinn: **+0.2 DQ**.
3. **Militärische Motivation präzisieren.** Im Brief „Rückkehr in ein vertrautes Umfeld“ zu „Rückkehr in ein mir vertrautes militärisches Umfeld“ ändern. Das verhindert, dass Bundeswehr und Schweizer Armee als dasselbe Umfeld gelesen werden. Erwarteter Gewinn: **+0.2 DQ**.
### Tier 3
1. Einen der unmittelbar aufeinanderfolgenden `Entwickelte`-Einstiege bei Fraunhofer/Vizrt variieren, ohne den Ownership-Grad zu erhöhen.
2. Die Wiederholung von „wesentlicher Grund“ und „wesentlicher Reiz“ in zwei benachbarten Briefabsätzen glätten.
3. Generali nur dann weiter kürzen, wenn zusätzlicher Platz benötigt wird. Der Eintrag hält die DACH-Chronologie nachvollziehbar und ist nicht schädlich.
## 9. Cover-Letter Decision und Kritik
**Decision: YES.** Das Motivationsschreiben erhöht den Wert des Pakets. Es ergänzt drei Dinge, die der Lebenslauf allein nicht leisten kann: die militärisch begründete Motivation, das persönliche Interesse am gesamten Stack und die praktische Passung mit Region Bern, B-Bewilligung und kurzem Arbeitsweg.
Die technische Argumentation ist spezifisch für diese Stelle und bleibt evidenzsicher. Die Offizierslaufbahn wird weder zu Führungserfolgen noch zu Nachrichtendienst- oder Cybererfahrung aufgeblasen. Das Schreiben ist mit 300 Haupttextwörtern am oberen Rand der Normalspanne, aber auf einer gut lesbaren Seite. Löschen würde das Paket schwächen.
Sprachlich ist der Brief deutlich natürlicher als vor Edit 2. Verbleibend sind nur die Tier-2-Präzisierung des „vertrauten Umfelds“ und kleine Wiederholungen. Keine Em-Dashes, kein generischer Unternehmenslob-Absatz und keine defensive Gap-Liste.
## 10. Mechanical Verification
- Kanonisches System und beide Dokumentvalidatoren: **PASS, 0 Warnungen**
- LaTeX: Lebenslauf **2 Seiten**, Motivationsschreiben **1 Seite**
- Logs: **0 overfull/underfull boxes**, keine LaTeX-Warnungen
- PDF: A4, keine Überlappung, keine abgeschnittenen Zeilen, keine verwaisten Überschriften
- Seitenzahlen: 1/2 und 2/2 vollständig innerhalb der Seitenkoordinaten
- `pdftotext -layout`: Arbeitgeber, Titel, Orte und Zeiträume in korrekter Reihenfolge
- Visuelle Prüfung: alle drei Seiten vollständig kontrolliert; Hierarchie, Abstände und Lesbarkeit sauber
- Unicode-Prüfung: keine En-, Em- oder geschützten Dashes in den Quellen; LaTeX-Doppelhyphen nur für Datumsbereiche
- AI-Fingerprint: keine auffälligen Gedankenstrichmuster oder generische Firmensprache; leichte Listen- und Verbkadenz bleibt ein kleiner Stilpunkt
## Schlussurteil
**Rollenfit:** 79/100, Core, Hard Gate PASS.
**Dokumentqualität:** 92/100.
**Channel:** Weak.
Die inhaltliche Positionierung ist jetzt deutlich stärker: Die Bundeswehr-Laufbahn ist an den beiden richtigen Stellen sichtbar und bleibt zugleich beweissicher. Vor einer Einreichung ist genau ein inhaltlicher Fix erforderlich, danach sind nur noch optionale Tier-2-Politur und die telefonische Klärung von Lohnklasse und Level offen.
output/KdoCy_DevOps_Data_Platform/e2e_kdocy_devops_data_platform_cover_letter.tex
+55
View File
@@ -0,0 +1,55 @@
% Kdo Cy: DevOps Engineer III (Data Platform). German-language Swiss/DACH letter.
% Every claim traces to a resume bullet and a canonical ID. The second-Amtssprache point is
% deliberately NOT raised: the user confirmed it is not a hard requirement, so naming it here
% would hand the reader an objection, which was the Aker BP CL-A finding.
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[ngerman]{babel}
\usepackage{lmodern}
\usepackage[a4paper,left=0.85in,right=0.85in,top=0.7in,bottom=0.7in]{geometry}
\usepackage{parskip}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\begin{document}
{\Large\bfseries Dennis Thiessen, M.Eng.}\par
Bern, Schweiz $\vert$
\href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$
\href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}
\vspace{1.5em}
Schweizer Armee, Kommando Cyber\\
Herr Marcel Matthey-Doret\\
Eichacher, 3086 Wald
\vspace{1em}
\today
\vspace{1em}
\textbf{Bewerbung als DevOps Engineer III (Data Platform), Ref. JRQ\$540-19848}
\vspace{0.8em}
Sehr geehrter Herr Matthey-Doret
Ihre Ausschreibung verbindet zwei Aufgaben, die mich fachlich besonders interessieren: den Betrieb einer Datenplattform mit MLOps-Umgebungen bis zur GPGPU-Hardware und die Administration der darunterliegenden Linux-Systeme. Bei Bosch arbeitete ich genau an dieser Schnittstelle. In der durchgehend laufenden Halbleiterfertigung waren robuste Automatisierung und klare Betriebsabläufe Voraussetzung. Dort integrierte ich containerisierte ML-Inferenz und administrierte die Linux-Systeme für die zugehörigen Workloads.
Die Infrastrukturautomatisierung habe ich bei Bosch nicht nur eingesetzt, sondern erweitert. Die Ansible-Playbooks lagen in Git und wurden über Jenkins- und GitLab-Pipelines ausgeführt. Für wiederkehrende Engpässe entwickelte ich eigene Ansible-Erweiterungen, darunter ein Plugin, das Zugangsdaten aus einem Passwort-Keystore lokal zwischenspeichert und dadurch wiederholte Abrufe reduziert. Heute verantworte ich bei Swisscom als Component Owner geschäftskritische Datenstrecken mit Oracle, Kafka und Teradata, einschliesslich Pikettdienst und Datenqualität. Für mich gehört der Betrieb damit zur Entwicklung: Erst im laufenden Einsatz zeigt sich, ob eine Plattform trägt.
Die Aufgabe hat für mich auch persönlich Gewicht. Ich diente sechs Jahre in der Bundeswehr, absolvierte den Offizieranwärterlehrgang und die Offizierschule und schied als Leutnant aus. Für die Schweizer Armee zu arbeiten wäre für mich eine Rückkehr in ein vertrautes Umfeld, nun mit meiner technischen Berufserfahrung. Zum Schutz der Schweiz und ihrer Bevölkerung beizutragen, ist ein wesentlicher Grund für meine Bewerbung.
Die Arbeit über den gesamten Stack bis in den Serverraum ist für mich ein wesentlicher Reiz der Stelle. Seit rund zehn Jahren betreibe ich einen eigenen Debian-Server mit nginx, Mailserver, VPN und Docker-Diensten, samt Härtung und Updates, weil mich diese Arbeit interessiert. Ich bin deutscher Staatsangehöriger mit Schweizer B-Bewilligung. Deutsch ist meine Muttersprache, und ich lebe in der Region Bern. Zimmerwald wäre für mich ein kurzer Arbeitsweg und eine langfristige Perspektive.
Ich freue mich auf die Gelegenheit zu einem Gespräch und darauf, mehr über die Rolle der Plattform im Zielbild Ihres Teams zu erfahren.
Freundliche Grüsse
\vspace{1.5em}
Dennis Thiessen
\end{document}
output/KdoCy_DevOps_Data_Platform/e2e_kdocy_devops_data_platform_resume.tex
+93
View File
@@ -0,0 +1,93 @@
% Kommando Cyber (Kdo Cy): DevOps Engineer III (Data Platform), Zimmerwald
% German-language Swiss/DACH profile. Position titles stay English (user decision 2026-08-27).
% Generated from claims.json only. New evidence this round: BS-6 (Linux estate + Ansible
% extensions) and PP-3 (self-hosted Debian). The pull-reconciliation buzzword the JD uses is
% forbidden by claims.json because execution was push-based from Jenkins/GitLab, so the Bosch bullet
% states the mechanism literally instead.
\documentclass{resume}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[ngerman]{babel}
\usepackage{lmodern}
\usepackage[a4paper,left=0.65in,right=0.65in,top=0.55in,bottom=0.55in]{geometry}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\usepackage[version=4]{mhchem}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[R]{\small \thepage/\pageref{LastPage}}
\name{Dennis Thiessen, M.Eng.}
\headline{Staff Data \& Platform Engineer $\vert$ Linux, Ansible, Kubernetes $\vert$ Datenplattformen, Produktionsbetrieb und ML-Deployment}
\contactline{Bern, Schweiz $\vert$ \href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$ \href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}}
\begin{document}
\begin{rSection}{Kurzprofil}
Staff Engineer mit über zehn Jahren technischer Berufserfahrung in Telekommunikation, Halbleiterfertigung, Broadcast-Technologie und Versicherung sowie sechsjähriger Offizierslaufbahn bei der Bundeswehr. Verantwortet bei Swisscom geschäftskritische Datenstrecken und Kubernetes-Anwendungen im Produktivbetrieb. Administrierte und automatisierte zuvor bei Bosch Linux-Systeme für Analyse- und ML-Workloads und integrierte dort containerisierte ML-Inferenz in eine durchgehend laufende Halbleiterfertigung.
\end{rSection}
\begin{rSection}{Kenntnisse}
\skillline{Programmiersprachen}{Python, SQL; Java und C\# aus früheren beruflichen Projekten}
\skillline{Linux und Automatisierung}{Linux-Administration (Bosch), Systemhärtung (laufender Eigenbetrieb), Ansible einschliesslich eigener Erweiterungen, Jenkins, GitLab CI/CD, CloudFormation (IaC)}
\skillline{Container, Cloud und MLOps}{Docker, Kubernetes, Microservices, MLOps (Deployment und Orchestrierung containerisierter ML-Inferenz), AWS (S3, Glue, Athena/Iceberg, Redshift)}
\skillline{Daten und Observability}{Apache Kafka, Airflow, PySpark, Oracle, Teradata, Hadoop/Impala, Data Products und Metadaten; ELK, Grafana, Prometheus, Loki}
\skillline{Zertifizierungen}{AWS Certified Solutions Architect (Associate); Data Engineering with AWS; iSAQB CPSA-F; ITIL Foundation}
\skillline{Sprachen}{Deutsch (Muttersprache), Englisch (fliessend), Norwegisch und Russisch (Grundkenntnisse)}
\end{rSection}
\begin{rSection}{Berufserfahrung}
\begin{rSubsection}{Swisscom (Schweiz) AG}{Okt. 2023 -- heute}{Staff Data, Analytics \& AI Engineer (seit Apr. 2025; zuvor Senior)}{Bern, Schweiz}
\item Verantworte als Component Owner geschäftskritische ETL-Strecken der Fulfillment-Domäne mit Oracle, Kafka, Python und Teradata, einschliesslich Produktivbetrieb, Datenqualität, Governance, Störungsbearbeitung und Pikettdienst.
\item Entwickle und betreibe Python-Datenanwendungen auf Kubernetes mit GitLab CI/CD, vom Container-Build über das Deployment bis zum laufenden Betrieb.
\item Migrierte die Pipelines der Domänen Fulfillment und Product Analysis auf die AWS-Plattform (Glue, Athena mit Apache Iceberg, Redshift, Airflow und CloudFormation als Infrastructure as Code, IaC) und unterstützte damit das übergreifende Migrationsprogramm.
\item Modelliere governance-konforme Data Products und pflege deren aktive Metadaten im unternehmensweiten Data Mesh, damit nachgelagerte Teams die Daten auffinden und einordnen können.
\end{rSubsection}
\begin{rSubsection}{Robert Bosch Semiconductor Manufacturing Dresden GmbH}{Feb. 2020 -- Dez. 2022}{Senior Engineer, Data Analysis (Data \& ML Engineering)}{Dresden, Deutschland}
\item Integrierte containerisierte ML-Inferenz mit Docker, Kubernetes und Ansible in eine durchgehend laufende Halbleiterfertigung und ermöglichte damit die automatisierte bildbasierte Defektklassifikation ohne manuellen Eingriff.
\item Administrierte und automatisierte die Linux-Serverlandschaft für Analyse- und ML-Workloads, darunter ML-Plattform, Docker-Hosts, Backend-Dienste und Ansible-Steuerung. Die Playbooks lagen in Git und wurden über Jenkins- und GitLab-Pipelines ausgeführt.
\item Erweiterte Ansible unter anderem um ein Plugin für Zugangsdaten, das Secrets aus einem Passwort-Keystore abruft und lokal zwischenspeichert. Dies reduzierte wiederholte Keystore-Abfragen und verbesserte Performance und Skalierbarkeit.
\item Implementierte einen Proof of Concept zur Anomalieerkennung mit ELK, Kafka und Docker sowie Monitoring, Logging und Alarmierung mit Grafana, Prometheus und Loki.
\item Entwickelte Datenservices in Python, Java und C\# auf Basis von Oracle und Hadoop/Impala für Defect Management und Prozessanalyse.
\item Verantwortete als Application Owner Analyseanwendungen und vorgelagerte Pipelines, einschliesslich SLOs, Anwenderschulung, technischer Dokumentation und Abstimmung mit Lieferanten.
\end{rSubsection}
\begin{rSubsection}{Fraunhofer CML}{Sep. 2018 -- Okt. 2019}{Research Software Engineer}{Hamburg, Deutschland}
\item Entwickelte containerisierte Microservices für eine maritime Forschungsplattform und führte eigenständig eine Jenkins-CI/CD-Strecke mit Quality Gates ein.
\end{rSubsection}
\begin{rSubsection}{Vizrt}{Jul. 2017 -- Mai 2018}{Test Automation / DevOps Engineer}{Bergen, Norwegen}
\item Entwickelte Python- und C++-Backend-Komponenten für ein verteiltes Video-Transcoding-System. Integrierte automatisierte Audio-/Video-Tests als Quality Gates in die CI/CD-Pipeline.
\end{rSubsection}
\begin{rSubsection}{Generali Deutschland Informatik Services GmbH}{Mai 2015 -- Jun. 2017}{IT Consultant}{Hamburg, Deutschland}
\item Führte BDD-Testautomatisierung mit einem Proof of Concept ein und übernahm die technische Verantwortung für Testsuite und Jenkins-Jobs.
\end{rSubsection}
\begin{rSubsection}{Bundeswehr}{Juli 2008 -- Nov. 2014}{Offizieranwärter und Offizier (zuletzt Leutnant)}{Deutschland}
\item Diente sechs Jahre in der Bundeswehr, absolvierte den Offizieranwärterlehrgang und die Offizierschule und schied als Leutnant aus.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Projekte}
\begin{rSubsection}{Privater Debian-Server}{seit rund zehn Jahren}{Eigenbetrieb}{}
\item Betreibe und administriere einen selbst gehosteten Debian-Server mit nginx, Mailserver, Nextcloud, VPN, Docker-Diensten und Bitwarden, inklusive Systemhärtung und Security-Konfiguration.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Ausbildung}
\compactentry{M.Eng. Computer Aided Engineering}{Apr. 2012 -- Okt. 2013}
Universität der Bundeswehr München, Schwerpunkt Software Design \& Engineering. Masterarbeit an der Tongji-Universität, Shanghai: \textit{Development of a Web-Based Remote Fault Diagnosis System}.
\compactentry{B.Eng. Information and Telecommunication Technologies}{Okt. 2009 -- Okt. 2012}
Universität der Bundeswehr München.
\end{rSection}
\end{document}
output/KdoCy_DevOps_Data_Platform/resume.cls
+59
View File
@@ -0,0 +1,59 @@
\ProvidesClass{resume}[2026/07/27 Evidence-first professional resume]
\LoadClass[11pt,a4paper]{article}
\RequirePackage{enumitem}
\RequirePackage{ifthen}
\RequirePackage{lastpage}
\RequirePackage{parskip}
\RequirePackage{needspace}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\def\@resumename{}
\def\@resumeheadline{}
\def\@resumecontact{}
\newcommand{\name}[1]{\def\@resumename{#1}}
\newcommand{\headline}[1]{\def\@resumeheadline{#1}}
\newcommand{\contactline}[1]{\def\@resumecontact{#1}}
\newcommand{\printresumeheader}{%
{\LARGE\bfseries \@resumename}\par
\vspace{0.15em}
\ifthenelse{\equal{\@resumeheadline}{}}{}{\@resumeheadline\par}
\vspace{0.1em}
{\small \@resumecontact}\par
\vspace{0.45em}
}
\AtBeginDocument{\printresumeheader}
\newenvironment{rSection}[1]{%
\vspace{0.35em}
{\bfseries #1}\par
\vspace{0.1em}\hrule\vspace{0.35em}
}{%
\vspace{0.2em}
}
% Arguments: employer, dates, formal title, location.
\newenvironment{rSubsection}[4]{%
\Needspace{6\baselineskip}%
{\bfseries #1}\hfill{\small #2}\par
{\itshape #3}\hfill{\itshape #4}\par
\vspace{0.1em}
\begin{itemize}[leftmargin=1.25em,label=\textbullet,itemsep=0.18em,topsep=0.15em,parsep=0pt,partopsep=0pt]
}{%
\end{itemize}
\vspace{0.25em}
}
\newcommand{\skillline}[2]{%
\textbf{#1:} #2\par
\vspace{0.08em}
}
\newcommand{\compactentry}[2]{%
\textbf{#1}\hfill #2\par
}
output/KdoCy_DevOps_Data_Platform/session_kdocy_devops_data_platform.md
+278
View File
@@ -0,0 +1,278 @@
# Session: Schweizer Armee — Kommando Cyber (Kdo Cy) — DevOps Engineer III (Data Platform)
## JD Integrity
- **File/source:** `JD_kdocy_devops_data_platform.txt` (this folder) — verbatim body of `https://jobs.admin.ch/offene-stellen/devops-engineer-iii-data-platform/752ded37-1753-4ac8-946a-b0c958d4424a`
- **Retrieval method and date:** direct HTTP fetch of the Prospective-ATS detail page, HTML stripped, **2026-08-27**. Not reconstructed, not paraphrased.
- **Verbatim posting:** YES
- **Posting status:** LIVE 2026-08-27. Confirmed present in the portal's own `Kdo Cy` search result set the same day (11 Kdo Cy roles open).
- **Source of the lead:** user, manually — `jobs.admin.ch` is **not** in the job_scout roster.
## JD Info
| Field | Value |
|---|---|
| Employer | Schweizer Armee — Kommando Cyber (Kdo Cy) |
| Role | DevOps Engineer III (Data Platform) |
| Reference | JRQ$540-19848 |
| Arbeitsort | **Eichacher, 3086 Wald (Zimmerwald)** |
| Pensum | 80–100%, Jobsharing möglich |
| Anstellungsart | unbefristet |
| Eintritt | nach Vereinbarung |
| Deadline | none stated |
| Contact | **Marcel Matthey-Doret, Chef Centers of Excellence CEA, +41 58 46 78409** |
| Language of posting | German |
## Application Decision
- **Audience profile:** Swiss/DACH (German-language federal employer, German-language posting)
- **Evidence Fit: 79/100 — Core (lower end)**. Trajectory, all on 2026-08-27: **55** as posted → **64** after the language/citizenship answers → **68** after the Ansible correction → **79** after the Linux-administration and Ansible-extension evidence. **The movement is KB repair, not re-framing** — two canonical blind spots (Ansible, Linux) were suppressing a real profile.
- **Fit class: Core** (lower end, 75+)
- **Hard gate: PASS.** All three original §1 triggers now cleared — see Gate Audit. Remaining honest weakness is **physical datacenter work (R6), a genuine Gap** that interest does not substitute for.
- **Channel Strength:** Weak today; **upgradeable to Moderate** — the posting publishes a named contact with a direct number
- **Channel plan:** phone Matthey-Doret **before** any document work — the same call resolves both gates
- **Cohort slot:** recorded on submission in evidence-first-2026-01, now **5/10 applications and Core 3/7**. Adjacent remains 2/2 and Stretch 0/1.
- **Decision: SUBMITTED 2026-08-27.** Evidence Fit 79/Core, hard gate PASS, Document Quality 92, channel Weak.
## Gate Audit (`critique_framework.md` §1)
**UPDATED 2026-08-27 — user-confirmed that neither the second Amtssprache nor Swiss citizenship is a hard requirement.** Two of three triggers cleared; one stands.
- ✅ **CLEARED — second Amtssprache.** Confirmed a Wunschkriterium, not a filter. It stays a competitive disadvantage against a candidate who has French, and it is a plausible interview topic, but it no longer gates the application.
- ✅ **CLEARED — citizenship.** A German citizen with a B permit is eligible. The PSP itself still happens; its level and timing are simply not disqualifying facts.
- ✅ **CLEARED 2026-08-27 — the title-defining capability.** This was called a Gap on an incomplete KB. It was wrong. The role runs the Linux/IKT layer beneath a data platform, and the user in fact has: **~3 years of professional Linux estate administration and automation at Bosch** (ML platform hosts, Docker hosts, backend services, Ansible control — BS-6), **Ansible extension development** (a credential plugin with local caching to cut keystore lookups — infrastructure engineering, not tool usage), and **~10 years of continuous self-hosted Debian administration and hardening** (PP-3). The operating layer is **Adjacent-to-Direct**, not a Gap.
- ⚠️ **REMAINING, and not to be papered over — R6 physical datacenter work.** Zero experience. The user states genuine interest, which honestly addresses the JD's *"Freude an der Arbeit über den gesamten Stack"* criterion — but interest is not experience, and the resume must not imply otherwise. This belongs in the cover letter as motivation, never in a bullet as evidence.
### Original audit, retained for the record
Three of the four NO-GO triggers fired as posted:
1. **A required qualification is a Gap** — *"Aktive Kenntnisse einer zweiten Amtssprache"*. German is his native tongue and is itself an Amtssprache; the requirement is for a **second** one (French, Italian, Romansh). `claims.json` `identity.languages` = German native, English fluent, **Norwegian basic, Russian basic**. Neither is an Amtssprache. There is no bridge and nothing to write around. **Note:** `config.md` records that the Swisscom Zeugnis lists French and Italian as HR boilerplate and that this is **explicitly forbidden** on any document — so the one place those languages appear on paper is a known falsehood and must not be used here of all places.
2. **The title-defining capability is not Direct** — the title is *DevOps Engineer (Data Platform)*, and the function as written is running the **IKT/Linux layer beneath** a data platform: whole-system Linux administration, performance tuning, hardening, plus physical datacenter duty. Dennis builds and owns **data products and pipelines on top of** managed platforms. Same authoring-vs-operating split as Aker BP, one layer lower.
3. **Clearance is unresolved** — the posting requires the ability to integrate into a *nachrichtendienstliches Umfeld*, is silent on nationality, and he is a German citizen. Unresolved either way.
## Requirements
| # | Requirement (verbatim sense) | Req/Pref | Class | Canonical evidence | Gate? |
|---|---|---|---|---|---|
| Q1 | IT-Abschluss HF/FH + mehrere Jahre Praxis im Engineering und Betrieb komplexer IT-Systeme | Required | **Direct** | EDU-MENG (M.Eng., exceeds HF/FH); ~13 yrs across SWISSCOM/BOSCH/FRAUNHOFER/VIZRT | — |
| Q2 | Fundierte DevOps-Kompetenz: Aufbau/Betrieb/Automatisierung **moderner Linux-Infrastrukturen**, umfassende Kenntnisse **IaC und GitOps** | Required | **Adjacent, upgraded 2026-08-27** | **IaC is Direct:** intensive Ansible work at Bosch plus production-current CloudFormation. **Linux administration is Direct via BS-6** and current personal hardening is PP-3. GitOps was clarified as Git-versioned, push-executed automation without a reconciliation agent; the named practice remains a Gap and the word stays forbidden. | ⚠ |
| Q3 | Lernbereitschaft, Offenheit für Nicht-Mainstream-Ansätze (Dhall, Pull-GitOps, Dew-/Fog-Computing) | Required (attitude) | **Neutral** | Attitude criterion; the named technologies are exotic and no candidate will hold them | — |
| Q4 | Hands-on über den ganzen Stack — **von physischer Infrastruktur im Rechenzentrum** bis Big-Data-Umgebungen **wie ClickHouse und Kafka** | Required | **Split** | **Kafka Direct** (production-current). ClickHouse: Gap but substitutable — BS-2 Oracle + Hadoop/Impala, SQL production-current. **Physical DC/server/network hardware: Gap, 0 evidence** | ⚠ |
| Q5 | Tadelloser Leumund + Fähigkeit, sich in ein **nachrichtendienstliches Umfeld** zu integrieren — *"werden vorausgesetzt"* | Required | **Constraint — RESOLVED 2026-08-27** | Swiss citizenship confirmed not a hard requirement (user). German citizen + B permit is eligible; a PSP still runs | ✅ |
| Q6 | **Aktive Kenntnisse einer zweiten Amtssprache** + gute Englischkenntnisse | Required | **Soft — RESOLVED 2026-08-27** | Confirmed not a hard requirement (user). English half Direct. Remains a competitive disadvantage and a likely interview question | ✅ |
| R1 | IaC/GitOps planen, implementieren, betreiben; ganzen Deployment-Zyklus verantworten | Core resp. | **Adjacent** | SW-3, FC-1, VZ-2, CloudFormation and BS-6. The Ansible flow was version-controlled and CI-driven but push-based, not Pull-GitOps. | — |
| R2 | **MLOps-Umgebungen** für Deployment, Monitoring, Betrieb von KI-/ML-Modellen **bis zur GPGPU-Hardware** | Core resp. | **Adjacent — strongest hook** | BS-1 (containerized ML inference integrated into a 24/7 semiconductor production environment) + SW-3. **GPGPU: Gap.** BS-1 forbids claiming model training or full-lifecycle ownership | — |
| R3 | Komplexe **Linux-Systemlandschaften** für den E2E-Betrieb der IKT-Schicht einer **Datenplattform** | Core resp. | **Adjacent-to-Direct** | Data-platform operation is Direct through SW-1, SW-2 and SW-7; professional Linux-estate administration and automation is Direct through BS-6. The two layers are proven in different roles rather than as one current end-to-end platform. | — |
| R4 | Linux ganzheitlich administrieren, Performance-Optimierung, **System-Hardening**, Security-Konfigurationen | Core resp. | **Adjacent-to-Direct — upgraded 2026-08-27** | **BS-6** (Bosch Linux estate: ML platform, Docker hosts, backend, Ansible control) + **PP-3** (~10 yrs self-hosted Debian: nginx, mail, Nextcloud, VPN, Docker, Bitwarden — administered and hardened). **Hardening evidence leans personal — state it as such**, never imply an employer mandate | — |
| R5 | Monitoring-, Logging-, Observability-Lösungen betreiben und weiterentwickeln | Core resp. | **Direct-to-Adjacent** | BS-4 (ELK/Kafka anomaly-detection PoC + monitoring) — scoped as a PoC, **never** an enterprise observability platform | — |
| R6 | Physische Infrastruktur im Rechenzentrum betreuen, Installationen und Wartung unterstützen | Core resp. | **Gap — remains** | 0 experience. User states genuine interest, which answers Q4's *Freude*-criterion honestly. **Cover letter as motivation; never a resume bullet** | ⚠ |
## Evidence Fit Breakdown
| Dimension | Weight | Score | Reasoning |
|---|---:|---:|---|
| Required qualifications | 35 | **30** | Q1 Direct, Q3 neutral, English Direct, **Q5 and Q6 resolved 2026-08-27**; Q4 remains split (physical DC absent); **Q2 improved — IaC is now Direct via Ansible + CloudFormation**, GitOps-as-practice and Linux depth still thin |
| Core responsibilities | 25 | **19** | R2/R3/R4/R5 are Direct-to-Adjacent and R1 is Adjacent; **R6 physical datacenter work remains a Gap**. GPGPU and Pull-GitOps also remain unevidenced. |
| Level and ownership | 15 | **10** | Engineer III vs current Staff/Engineer IV; Component Owner and Application Owner scope (SW-2, BS-3) comfortably covers the ownership asked. Reads lateral-to-slightly-below |
| Recency and depth | 10 | **9** | Kubernetes, Kafka, Python, SQL, CI/CD all production-current. BS-1/BS-4 are the strongest role-specific proofs but date to Bosch, not today |
| Domain/tool transfer | 10 | **7** | ClickHouse ← Impala/Oracle/SQL: yes. Git-versioned CI/CD automation transfers partly to GitOps concepts but is not Pull-GitOps. **Physical datacenter hardware and GPGPU are not substitutable.** Linux administration and hardening are now directly evidenced through BS-6 and PP-3. |
| Practical constraints | 5 | **4** | Best commute in the entire log, unbefristet, 80–100%, language and citizenship both cleared. Held back only by **comp — the Lohnklasse is still unknown** |
| **Total** | | **79** | **Core (lower end)** — 55 → 64 → 68 → **79**, every step a KB correction rather than a re-framing |
**Current fit state after the full 2026-08-27 KB repair:** language and citizenship constraints are cleared, and Linux/IKT operation is now evidenced through BS-6 plus PP-3. The remaining evidence gaps are physical datacenter work, Pull-GitOps, GPGPU and ClickHouse by name. The resulting score is 79/Core.
**Still open and unasked: compensation.** Bund Lohnklasse for a DevOps Engineer III is not public and must be asked, not researched — the same class as the SBB Anforderungsniveau K. Current comp is ~140k; the local tier permits near-current, not a steep cut.
## ATS Keywords
- **Direct, usable:** DevOps, CI/CD, GitLab CI/CD, Kubernetes, Docker, Container, Kafka, Python, SQL, Datenplattform, Data Platform, Data Pipelines, Monitoring, Logging, Observability, ELK, Automatisierung, Infrastructure as Code, CloudFormation, Deployment, AWS
- **Bridge, use with care:** MLOps, ML-Inferenz im Produktivbetrieb, Betrieb 24/7, Datenqualität, Governance, On-Call, Incident
- **Gap — do NOT seed:** GitOps, Dhall, Pull-GitOps, Dew-/Fog-Computing, ClickHouse, GPGPU, Rechenzentrumsbetrieb, Terraform, Zero Trust
- **Forbidden here by `claims.json`/`config.md`:** French, Italian (Zeugnis boilerplate — a known falsehood), LangChain/LangGraph/LlamaIndex, Azure, GCP, Terraform, "built the observability platform", "own the data platform"
## Gap Assessment
1. **Physical datacenter work** — zero experience and a genuine downward-scope signal for someone at Staff level.
2. **Pull-GitOps as a named practice** — the real flow was Git-versioned but push-executed from Jenkins/GitLab without a reconciliation agent.
3. **ClickHouse and GPGPU** — Impala/Oracle/SQL and containerized inference provide honest bridges, not Direct evidence.
4. **Second Amtssprache** — confirmed as a soft criterion rather than a hard gate; still a competitive disadvantage.
5. **PSP** — eligibility is cleared, but no present or former clearance is claimed and the normal screening process remains.
6. **Level and compensation** — Engineer III may be lateral or slightly below current Staff scope; Lohnklasse remains unknown.
## Company Context
- **Kommando Cyber (Kdo Cy)** — stood up **1 Jan 2024** as the Army's cyber/ICT command; ~9,000 civilian and military staff across the Army, Kdo Cy itself being the ICT and cyber-defence arm for the Army and the Sicherheitsverbund Schweiz.
- **Neue Digitalisierungsplattform (NDP)** — Kdo Cy's flagship programme: a secure, robust, resilient platform as the technical foundation for digitising the Army, consolidating today's decentralised datacenter infrastructure; first services expected from **July 2026**. This role's "Datenplattform" sits in that world and is the single best-informed thing to raise on the call.
- **Zimmerwald is the Zentrum elektronische Operationen (ZEO).** This is the Army's SIGINT evaluation site: intercepted satellite and cable communications collected elsewhere are evaluated there, and Comint reporting goes to the NDB. That is exactly what the JD's *"erfasste zivile und militärische Kommunikationsdaten … Erkenntnisse unterstützen den NDB und den MND"* describes.
- **Material implication for the user, not just for the documents:** this is an intelligence job, not a corporate platform job. It carries a clearance process, standing public and parliamentary scrutiny of Funk-/Kabelaufklärung, and a career narrative that is harder to reverse than a normal move. Worth wanting deliberately.
## Level Read — user raised overqualification 2026-08-27
**Question:** the JD asks only for *"Abgeschlossene Ausbildung im IT-Bereich HF/FH"*, not a university degree — is the seat below his level?
- **The education line is the weakest of the level signals.** Swiss postings state a *minimum*; exceeding it is routine and nobody is filtered out for holding an M.Eng. Do not read the band off this line.
- **"III" is Senior, and that is primary-source, not inference.** The same unit's parallel posting on the same portal reads **"ICT-Architekt/-in III (Senior ICT Systems Architect)"** — Kdo Cy glosses its own roman numeral as Senior. So III is not a junior band.
- **The responsibility language is nonetheless practitioner-level.** The role is *"fachlich mitverantwortlich"* — shared technical responsibility, not lead. No people leadership, no architecture ownership, and it carries **physical datacenter installation and maintenance support**. Against his current Staff/Engineer IV plus Component Owner (SW-2) and prior Application Owner (BS-3) scope, this reads **lateral at best, plausibly a half-step down**.
- **Same shape as two known cases:** BKW was **declined** on exactly this (lateral, no upside), and SBB was flagged *"lateral or below"* and submitted anyway. `user_role_targeting_energy_trading` records the standing rule: no lateral same-tier moves. The Bern/Thun local tier is the deliberate exception, and it is what this role trades on.
- **The sharper risk is the mirror image: being screened out as overqualified.** Public-sector HR does filter for this, particularly when a Lohnklasse cannot match current comp. That makes the pay question do double duty — it is both his decision input and their objection.
- **The posting names no Lohnklasse** (0 hits). Federal pay classes exist publicly, but which class this seat sits in is not stated and must be asked, not assumed — same discipline as the SBB Anforderungsniveau K.
- **No better-levelled technical seat exists at Kdo Cy today.** Of the 11 open roles, the alternatives are PM/portfolio track (Portfolio-Manager, Senior Projektleiter, Projektmanagement Officer — not his lane) or ICT-Architekt III, which is nominally higher but a worse content fit (network/security architecture) and carries a harder language bar (second Amtssprache *plus* a passive third).
**Post-submission option:** a concise call to Matthey-Doret could still clarify Lohnklasse/band, how Engineer III is levelled and the development path for someone arriving at Staff scope. No such conversation is recorded, so channel strength remains Weak.
## Competitive Read
- **Obvious-fit candidate:** a Swiss-national senior Linux/DevOps engineer out of a Bund contractor, RUAG, armasuisse or a Swiss ISP — Linux fleet operations in the bones, IaC/GitOps daily, holds or has held a PSP, speaks German plus French, and is comfortable in a rack.
- **Dennis's advantage:** the **data side**. Real Kafka in production, governed data products, pipeline ownership with SLOs and on-call (SW-2), containerized ML inference put into a 24/7 production environment (BS-1), ELK/Kafka anomaly detection (BS-4). A Linux/DevOps generalist will not have built the platform's *data* half. Plus native German, EU/B-permit work authorisation, and he lives ~a village away.
- **Their advantage:** they are Direct on the exact axis the title names — Linux, hardening, GitOps, datacenter — they clear the language requirement without a conversation, and their clearance path is uncomplicated.
- **Level/scope:** Engineer III against his Staff/Engineer IV. Not a step up.
## Framing Strategy *(only if the call clears the gates)*
- **Professional identity:** Staff Data & Platform Engineer who builds and operates production data platforms — containers, CI/CD, streaming, observability — and has taken ML inference into 24/7 production.
- **Lead narrative:** the data platform's operating half, not the rack's.
- **Strongest proof points:** SW-2 (Component Owner, business-critical ETL with data quality, governance, incidents, on-call), SW-3 (Python applications on Kubernetes with GitLab CI/CD), BS-1 (containerized ML inference into 24/7 semiconductor production — the honest MLOps hook), BS-4 (ELK/Kafka anomaly detection and monitoring), SW-7 (governed data products, scoped **inside** the company-wide Data Mesh — never as owner), Kafka.
- **Honest bridges:** IaC via **CloudFormation**, never implied as Terraform. GitOps approached from GitLab CI/CD, named as adjacent rather than held. ClickHouse approached from Oracle/Impala/SQL.
- **Explicit gaps to state rather than paper over:** physical datacenter work, Pull-GitOps as a practice, GPGPU and ClickHouse by name.
- **Scope Discipline watch:** SW-1 must stay scoped to his own domains' pipelines; SW-7 must never read as owning the Mesh; BS-4 must never become "the observability platform".
- **User directives:** include the Bundeswehr officer career as the defence-role USP; explain the motivation to return to a military environment and contribute to protecting Switzerland and its population; remove em-dash punctuation; remove awkward German; omit C++/JavaScript from the skills block.
- **GitOps clarification resolved:** playbooks lived in Git and were applied push-style from Jenkins/GitLab. No reconciliation agent was present. The word stays forbidden; describe the mechanism literally.
## Bullet Plan — Phase 1 (proposed 2026-08-27, awaiting confirmation)
**Confirmed by user 2026-08-27:** documents in **German**; bundle **ML/MLOps primary + Data Engineer secondary**; format 2-page Swiss/DACH.
Budget per `config.md`: **11–14 bullets**. Recommended set = **10**, leaving 1–4 discretionary slots.
### Swisscom — Staff Data, Analytics & AI Engineer (Oct 2023 – heute)
| | ID | Achievement | JD hook | Match |
|---|---|---|---|---|
| * | SW-3 | Python-Anwendungen auf Kubernetes, GitLab CI/CD, voller Deployment-Zyklus | R1 Deployment-Zyklus, Q2 DevOps | Direct |
| * | SW-2 | Component Owner, Oracle/**Kafka** → Teradata, Pikett, SLA, Data Governance | Q4 Kafka, R3 Betrieb, "tadelloser" Betriebsnachweis | Direct |
| * | SW-1 | Migration der eigenen Domänen auf AWS, **CloudFormation = IaC** (scope-corrected: seine Domänen, nie "die" Migration) | Q2 IaC | Direct (IaC) / Adjacent (rest) |
| * | SW-7 | Governed Data Products + aktives Metadatenmanagement im unternehmensweiten Data Mesh | R3 Datenplattform | Adjacent |
| o | SW-8 | Domänen-gegroundete LLM-Agenten (Modellauswahl + Wissensbasis) | R2 KI-Modelle — but configuration, not MLOps | Bridge |
| o | SW-5 | Security Champion 2025/2026 | R4 System-Hardening/Security-Konfigurationen — the JD *does* name security, which is the `config.md` trigger. Rotating team role, weak signal | Weak |
| x | SW-4 | B2B-Datenprodukte, Dashboards | wrong axis for an operating role | — |
| — | SW-6 | PySpark | fold into Skills, not a bullet | — |
### Bosch Halbleiterwerk Dresden — Data & ML Engineer (Feb 2020 – Dez 2022)
| | ID | Achievement | JD hook | Match |
|---|---|---|---|---|
| * | BS-1 | Containerisierte ML-Inferenz (Docker, Kubernetes, **Ansible**) in 24/7-Halbleiterfertigung | **R2 MLOps — the anchor bullet**, plus Q2 IaC via Ansible | Direct-to-Adjacent, strongest |
| * | BS-4 | Anomalieerkennung mit ELK + Kafka, Grafana/Prometheus/Loki | **R5 Observability — direct hit**; Kafka again | Direct |
| * | BS-2 | Datenservices über OracleDB und Hadoop/ImpalaSQL | **Q4 ClickHouse bridge** (columnar/Big-Data analytics store) | Bridge |
| * | BS-3 | Application Owner: SLOs, Schulung, Dokumentation, Vendor, 24/7-Betrieb | R3 E2E-Betrieb, Ownership at level | Direct |
| * | **BS-6** | **Linux-Serverlandschaft administriert und automatisiert** (ML-Plattform, Docker-Hosts, Backend, Ansible-Control); **Ansible-Erweiterungen entwickelt** — u.a. Credential-Plugin mit lokalem Caching gegen Keystore-Abrufe | **Q2 + R3 + R4 — the bullet that turns the operating layer from Gap to evidence.** Ansible-Erweiterung ist Infrastruktur-Engineering, nicht Toolnutzung | **Direct** |
| x | BS-5 | Spotfire + TAF 2022 | BI, wrong axis here | — |
### Fraunhofer CML — (Feb 2017 – Jan 2020)
| | ID | Achievement | JD hook | Match |
|---|---|---|---|---|
| * | FC-3 | Containerisierte Microservices (Docker) für die MISSION-Datenaustauschplattform | early Docker/microservice depth | Adjacent |
| o | FC-1 | Jenkins-CI/CD eigenständig eingeführt | Q2 automation initiative | Adjacent |
| x | FC-2 | ARTUS NLP/Spracherkennung | model work, not operations | — |
### Vizrt — (2015 – 2017)
| | ID | Achievement | JD hook | Match |
|---|---|---|---|---|
| * | VZ-1 | Python/C++ Backend für ein **verteiltes** Video-Transcoding-System | JD: *"grosse, verteilte Dateninfrastruktur"* — the only distributed-systems evidence outside data pipelines. Scope: contributed; broadcasters are product context, never delivery claims | Bridge |
| o | VZ-2 | Automatisierte A/V-Testsuite + Quality Gates in CI/CD | CI/CD depth | Adjacent |
### Projekte (neuer Abschnitt — empfohlen)
| | ID | Achievement | JD hook | Match |
|---|---|---|---|---|
| * | **PP-3** | **Eigener Debian-Server, seit ~10 Jahren selbst betrieben, administriert und gehärtet** — nginx, Mailserver, Nextcloud, VPN, Docker-Dienste, Bitwarden | **R4 Hardening/Security-Konfiguration — the only *current* and *continuous* Linux evidence** (Bosch ended 2022). Belegt zugleich Q3/Q4: echtes Interesse am ganzen Stack, nicht behauptet sondern seit einem Jahrzehnt gelebt | Direct (personal) |
| x | PP-1 | Investing-/Signal-Plattform | wrong domain here | — |
**PP-3 wording discipline:** never enterprise/production/multi-user, no uptime or scale figures, always visibly a personal self-hosted system. Framed that way it is a strength for this employer — a candidate who runs and hardens his own mail and VPN server is exactly the profile a `Dhall`/`Pull-GitOps` team recognises.
### Generali — omit entirely (LOW on every axis for this JD)
**Recommended total: 12** — SW-3, SW-2, SW-1, SW-7, BS-1, **BS-6**, BS-4, BS-2, BS-3, FC-3, VZ-1, plus **PP-3** in a Projekte section. **Inside the 11–14 budget**, 2 discretionary slots left. Fill order if wanted: FC-1 (CI/CD-Initiative), SW-8 (aktuelles KI-Signal), VZ-2. SW-5 now unnecessary — BS-6 and PP-3 carry the security/hardening axis far better than a rotating Security-Champion role.
**Not writable, by canonical rule:**
- **The word GitOps** — resolved 2026-08-27 and the answer was no: Git-versioned, but push-executed from Jenkins/GitLab, no reconciliation agent. Write the substance instead: *"versionskontrollierte Infrastrukturautomatisierung mit Ansible aus Git, ausgeführt über Jenkins-/GitLab-Pipelines."* Against a team that names **Pull-GitOps** in its own posting, that precision reads better than the buzzword.
- **Physical datacenter work** — no experience. Interest belongs in the cover letter, never in a bullet.
- GPGPU; ClickHouse by name; Terraform. Hardening: writable, but visibly sourced to PP-3.
**Phase 2 issues to settle before generation:**
1. **German templates do not exist yet.** Every prior package is English, including SBB's German-language JD. Section headings, date formats and the skills block all need German equivalents in the `.tex`.
2. **Position titles in a German document** — keep the official English titles, or render them in German? Bosch's own Zeugnis title is German.
3. Bosch title variant: `experience_bosch.md` sanctions "Data & ML Engineer" for ML-leaning JDs; that fits the MLOps framing here.
## Critique Context
- **Reviewer persona:** a Bund technical lead in a security-cleared ICT unit, reading in German, who will check Linux depth and operating experience first and data-platform breadth second.
- **Domain vocabulary:** Betrieb, IKT-Schicht, Systemlandschaft, Härtung, Verfügbarkeit, Lebenszyklus, Deployment-Zyklus, Datenplattform, Observability.
- **Likely first technical challenge:** *"Beschreibe eine Linux-Systemlandschaft, die du selbst betrieben und gehärtet hast."* He can answer with the professionally administered Bosch estate (BS-6) and must separate it clearly from the current personal Debian hardening evidence (PP-3).
- **Second challenge:** distinguish the Git-versioned, push-executed Ansible flow from Pull-GitOps; then address the physical-datacenter gap directly.
## Cover Letter Plan
- **Decision: YES — but only if the gates clear.** A Swiss federal application in German expects a Motivationsschreiben, and this role needs prose to do two things a resume cannot: explain why a data engineer is applying to an operating-layer role, and address the second-language question openly instead of letting the reader find it.
- **Language: German.** Non-negotiable for this employer.
- **Paragraph structure:** (1) NDP and the data platform, concretely — why this specific platform; (2) the data half he brings — Kafka, governed pipelines with on-call ownership, ML inference in 24/7 production; (3) honest positioning on the operating layer and the language, without a defensive gap list (`cl_reference.md` bans that — the Aker BP CL-A finding); (4) local, long-term, native German, EU/B permit.
- **Verified hooks:** Neue Digitalisierungsplattform and the datacenter consolidation; Kdo Cy's founding in Jan 2024; Zimmerwald/ZEO as the site. All from first-party or press sources dated in the session, none inferred.
- **Jargon level:** German technical, Bund register, no Anglicism padding.
## Status
- **Phase 0: DONE** — 2026-08-27
- **Fit gate: PASS.** Evidence Fit **79/100 (Core, lower end)**, hard gate **PASS**. The original language and citizenship questions are resolved; the physical-datacenter gap remains explicit.
- **Phase 1: DONE 2026-08-27** — plan confirmed by user. The original count was recorded incorrectly; the generated CV had 15 bullets, not 13.
- **Phase 2: DONE after Edit 2, 2026-08-27** — German resume and motivation letter edited, compiled and visually inspected.
- Language **German**, position titles **English** (user decision). First German-language package in the log; headings, dates and the skills block were built for it (`ngerman` babel added).
- **15 bullets:** Swisscom 4 (SW-2, SW-3, SW-1, SW-7) · Bosch 6 (BS-1, **BS-6** ×2, BS-4, BS-2, BS-3) · Fraunhofer 1 (FC-3+FC-1) · Vizrt 1 (VZ-1+VZ-2) · Generali 1 (GN-1) · **Bundeswehr 1 (BW-1)** · **Projekte 1 (PP-3)**.
- **User-directed content change:** the weak SW-8 LLM-configuration bullet was removed and replaced with the role-specific USP, the six-year Bundeswehr officer career. BW-1 is now canonical and appears in the summary and as a separate experience entry.
- **Deviation from the proposed plan, deliberate: Generali was re-included.** The plan said omit, but a DACH-convention CV should not carry an unexplained 2015–2017 gap. Capgemini stays omitted per the standing user preference.
- **Verification:** canonical and document validators **PASS, 0 warnings**; MiKTeX pdflatex **2-page CV + 1-page letter, 0 overfull/underfull boxes**; `pdftotext` extraction order correct; visual review clean; forbidden/requested-term sweep **0 hits**.
- **Punctuation:** no Unicode en/em dash and no rendered narrative dash punctuation remain in either source. LaTeX double hyphens remain only for date ranges.
- **Critique v3 is CURRENT, 2026-08-27.** Evidence Fit **79/Core**, hard gate **PASS**, Document Quality **92/100**, Channel **Weak**. The Bundeswehr evidence strengthens the competitive and motivation narrative but does not add technical, intelligence or clearance fit points.
- **Known accepted deviation at submission:** the Bosch BS-1 bullet says the automated classification ran *„ohne manuellen Eingriff“*. Canonical BS-1 supports a qualitative reduction in manual classification, not complete elimination. The user reviewed the critique, judged the package good enough and submitted without this proposed fix.
- **Grades omitted** per the `grade_output_rule` of 2026-08-26 (thesis 1.0 stays off the document by default; say the word to include it).
- **Package status: SUBMITTED 2026-08-27.** Await response.
- **Next action:** no document work. Await screening; an optional post-submission call to **Marcel Matthey-Doret, +41 58 46 78409** could clarify Lohnklasse, Engineer-III level and development path and would upgrade the channel only if an actual conversation occurs.
## Submission Record
- **Submitted:** 2026-08-27, user-confirmed.
- **Fit class / score:** Core, Evidence Fit 79/100.
- **Hard gate:** PASS.
- **Document Quality:** 92/100.
- **Channel at submission:** Weak.
- **Outcome:** applied / awaiting response.
- **Documents:** German 2-page CV and German 1-page motivation letter.
- **Known accepted deviation:** the user submitted without applying the critique v3 BS-1 wording correction from *„ohne manuellen Eingriff“* to a qualitative reduction claim.
- **Cohort:** evidence-first-2026-01, application 5/10; Core 3/7, Adjacent 2/2, Stretch 0/1.
## Output Files
- Session: `output/KdoCy_DevOps_Data_Platform/session_kdocy_devops_data_platform.md`
- Verbatim JD: `output/KdoCy_DevOps_Data_Platform/JD_kdocy_devops_data_platform.txt`
- Resume source: `output/KdoCy_DevOps_Data_Platform/e2e_kdocy_devops_data_platform_resume.tex`
- Resume PDF: `Dennis_Thiessen_Lebenslauf.pdf` (2 pages, German)
- Cover letter source: `e2e_kdocy_devops_data_platform_cover_letter.tex`
- Cover letter PDF: `Dennis_Thiessen_Motivationsschreiben.pdf` — **CURRENT after Edit 2**, German, 1 page, **300 body words**, validator PASS, 0 boxes. Addressed to Marcel Matthey-Doret by name.
- Critique: **CURRENT v3, 2026-08-27**, `critique_kdocy_devops_data_platform.md`. Evidence Fit **79/Core**, Document Quality **92/100**, Channel **Weak**, hard gate **PASS**. The remaining BS-1 wording finding was knowingly accepted by the user at submission.
## Cover Letter — decisions taken at write time (2026-08-27)
- **Hooks are 100% first-party from the verbatim posting** (the JD's own MLOps/GPGPU + Linux-layer combination, the reference number, the named contact). No external claim needed verification because none was used.
- **The NDP hook was dropped.** Kdo Cy's Neue Digitalisierungsplattform is verified as a real programme, but nothing establishes that *this* data platform is part of it. Asserting the link would have been an inference dressed as familiarity. `cl_reference.md`: omit a hook rather than spend a paragraph proving company familiarity.
- **The second-Amtssprache question is deliberately NOT raised — a reversal of the Phase 0 plan.** That plan was written while the requirement still looked like a gate. The user then confirmed it is not a hard requirement, so naming it in the letter would hand the reader an objection they had already set aside. This is exactly the Aker BP **CL-A** finding ("I have not authored an enterprise governance framework myself"). The letter states the language assets positively instead; the topic belongs in the interview if it comes up.
- **The physical-datacenter point is framed as motivation, never as evidence** — *"Dass die Stelle über den gesamten Stack geht, bis in den Serverraum, ist für mich ein Argument dafür"*, carried by PP-3, which is visibly personal.
- **Overqualification is answered implicitly, not raised explicitly** — the closing asks about the platform's role in the team's Zielbild, which invites the level conversation without conceding anything.
- Bundle §S5 of `bundle_ml_ai_engineer.md` was checked for the scope-violation pattern found in `bundle_data_engineer.md` on 2026-08-21. **Clean** — its hook matches BS-1's canonical scope.
## Critique v3 (2026-08-27, after Edit 2)
- **Evidence Fit:** 79/100, Core (lower end). **Hard gate:** PASS. The new BW-1 evidence materially improves military-context credibility and motivation but does not establish technical military work, intelligence experience or clearance, so the fit score remains unchanged.
- **Document Quality:** 92/100. **Channel:** Weak.
- **Tier 1:** BS-1 currently says the automated defect classification ran *„ohne manuellen Eingriff“*. Canonical BS-1 supports qualitative reduction of manual classification, not complete elimination. Safe direction: *„... ermöglichte damit die automatisierte bildbasierte Defektklassifikation und reduzierte den manuellen Klassifikationsaufwand.“*
- **Tier 2:** compress the five-line summary; change *„CloudFormation als Infrastructure as Code, IaC“* to *„CloudFormation für Infrastructure as Code (IaC)“*; specify *„mir vertrautes militärisches Umfeld“* in the letter.
- **Mechanical re-check:** validators PASS with 0 warnings; CV 2 pages, letter 1 page; 0 box warnings; A4; extraction order correct; all three pages visually clean; footer 2/2 inside page bounds; no Unicode dash punctuation in either source.
## Edit History
### Edit 1 (2026-08-27): three Tier 1 relevance fixes
- **Source:** `critique_kdocy_devops_data_platform.md`, Tier 1 items 1–3. User approved with "apply tier 1 fixes".
- **Changes, all MODIFY — no bullet added, removed or swapped, no claim, scope or hedged verb touched:**
1. SW-3: "von der containerisierten Auslieferung" → **"vom Container-Build über das Deployment"**. The JD's own phrase is *"den gesamten Deployment-Zyklus verantworten"*. Deliberately descriptive — *verantworte* was **not** used, because SW-3's canonical scope is "within the team environment" and an ownership verb here would have been an escalation.
2. BS-4: "für Monitoring und Alarmierung" → **"für Monitoring, Logging und Alarmierung"**. Logstash and Loki are the logging half; the bullet had been under-describing work already done.
3. Skills: "Container und Cloud" → **"Container, Cloud und MLOps"**, with "Deployment und Betrieb containerisierter ML-Inferenz (MLOps)". Labelled as the demonstrated work (BS-1 + BS-6), never as a generic MLOps toolchain.
- **Result:** JD-term coverage **64% → 69%** (27/39). *Deployment* and *Logging* now present; *MLOps* now in the CV, not only the letter.
- **Verification:** validator **PASS, 0 warnings** · **2 pages** · **0 overfull/underfull boxes** · forbidden-term sweep over the rendered PDF **0 hits** · visual check clean, no new orphan or header wrap.
- **Delta vs baseline:** pages 2→2, boxes 0→0, char violations 0→0, bullets 13→13, coverage 64%→69%.
- **Estimated post-edit Document Quality ~92/100** (Relevance 11→14.5 of 15). **This is an estimate, not a scored critique — never quote it as a score.** Re-run `/critique` if a current number is needed.
- **Not applied:** Tier 2 items 4–7 and Tier 3 items 8–10 remain open, including the Performance/Skalierbarkeit rationale on the Ansible-plugin bullet (Tier 2 #5) and the near-verbatim CV sentence in the cover letter (Tier 2 #7).
### Edit 2 (2026-08-27): German language pass, Tier 2 fixes and Bundeswehr USP
- **Source:** direct user feedback in this session.
- **User-requested language edits:** removed C++/JavaScript and the word *begrenzt* from Skills; changed *Baue und betreibe* to *Entwickle und betreibe*; rewrote the Vizrt bullet in technical language; removed awkward constructions throughout the CV and letter.
- **Dash punctuation:** removed all Unicode en/em dashes and all narrative LaTeX dash punctuation from both `.tex` files. LaTeX double hyphens remain only where they correctly render date ranges.
- **Tier 2:** spelled out *Infrastructure as Code* beside IaC; added the canonically supported Performance/Skalierbarkeit outcome to BS-6; named Swisscom in the summary; fully paraphrased the Bosch cover-letter evidence around the operating constraint.
- **Content selection:** added canonical employment record `BUNDESWEHR` and claim **BW-1** from the existing structured extraction and experience file. Replaced SW-8 with a dedicated Bundeswehr entry; surfaced the six-year officer career in the summary and in the letter's motivation. No clearance, NATO, combat, command-scope or technical military claim was added.
- **Cover letter:** now 300 body words and uses the officer career as the personal motivation for returning to a military environment and contributing to the protection of Switzerland and its population.
- **Verification:** canonical system PASS, 0 warnings; both document validators PASS; CV 2 pages, letter 1 page; 0 overfull/underfull boxes; `pdftotext` order correct; visual inspection clean; 15 bullets before and after content swap.
- **Critique status:** v2 is stale after this material edit. Evidence Fit remains 79/Core; Document Quality was not re-scored.
output/RUAG_AI_Engineer_C5I/JD_ruag_ai_engineer_c5i.txt
+70
View File
@@ -0,0 +1,70 @@
Source URL: https://jobs.ruag.ch/offene-stellen/ai-engineer-c5i/4afaa9aa-74c1-4b4d-9232-cf65651ce8ec
Apply URL: https://jobs.ruag.ch/umantis/2514/de/apply/18027
Retrieved: 2026-08-27
Status at retrieval: LIVE; application form reachable
Official portal listing: Berufserfahrene, Thun, Schweiz, 80-100%
AI Engineer C5I (w/m/d)
Rund 3000 Mitarbeitende von RUAG und RUAG Real Estate leisten jeden Tag einen wesentlichen Beitrag zur Sicherheit der Schweiz. Sie sorgen dafür, dass die Schweizer Armee sowie andere Einsatz- und Sicherheitsorganisationen ihre Aufgaben jederzeit umfassend wahrnehmen können.
Das kannst du bewegen
* Aufbau, Weiterentwicklung und Betrieb einer internen AI-/LLM-Plattform inkl. Compute Cluster
* Implementierung von Inhouse LLM-Lösungen inkl. Connectors, Tools und Retrieval-/Dokumentenlösungen
* Entwicklung von Datenpipelines, Data Preprocessing und Schnittstellen zu internen Systemen
* Umsetzung von Proof of Concepts (POCs), MVPs und Pilotlösungen im Bereich AI / R&D
* Containerisierung, Deployment und Automatisierung von AI-Workloads
* Technische Dokumentation sowie Wissenstransfer innerhalb des Teams
Das bringst du mit
* Hochschulabschluss (BSc/MSc) in Informatik, Computer Science, Data Science oder vergleichbar
* Mehrjährige Berufserfahrung in ML- Engineering, Data Engineering, Software Engineering oder DevOps
* Sehr gute Python-Kenntnisse sowie Erfahrung mit modernen ML-/AI- Frameworks
* Erfahrung mit Linux, Docker, Kubernetes oder vergleichbaren Plattformen
* Kenntnisse in LLMs, RAG, APIs, Data Pipelines oder Search-Technologien von Vorteil
* Strukturierte, selbständige und pragmatische Arbeitsweise mit hoher Umsetzungskompetenz
* Sehr gute Deutschkenntnisse in Wort und Schrift; Englischkenntnisse von Vorteil
* Militärische Karriere als höherer Unteroffizier oder Offizier von Vorteil
Erfüllst du nicht alle Voraussetzungen hundertprozentig? Bewirb dich trotzdem. Wir sind stets auf der Suche nach talentierten und motivierten Personen, die unser Team bereichern möchten. Wir legen grossen Wert auf Vielfalt und Chancengleichheit und unser Ziel ist es, ein vielfältiges Team aufzubauen, in dem alle ihre persönlichen Stärken voll entfalten können.
Wir ermutigen insbesondere alle Personen, unabhängig von ihrem Geschlecht, die möglicherweise Bedenken haben, eine Stelle in einem armeenahen Umfeld anzunehmen, sich zu bewerben.
Deine Ansprechperson
Frank Haugwitz
recruiting-c5i@ruag.ch
Arbeitsort
C5I, Uttigenstrasse 36, 3600 Thun
Über den Bereich
Die Business Area C5I der RUAG MRO Holding AG unterstützt die Schweizer Armee bei der Realisierung von IKT-Projekten mit spezifischem Fokus auf das Kommando Cyber. Dies ist deine Chance, einen Beitrag zu hochsicheren und souveränen IT-Systemen zu leisten. Wir sind in den letzten acht Jahren von einem kleinen «Start-up» mit null Mitarbeitenden auf über 260 Mitarbeitende gewachsen und bieten ein einmaliges, dynamisches Umfeld.
Hier kannst du an vorderster Front bei der Digitalisierung der Schweizer Armee mitarbeiten. Willst du einen Beitrag zum sicherheitspolitischen Umfeld leisten? Dann bist du bei uns genau richtig. Wir arbeiten agil und richten uns konsequent nach dem DevSecOps-Modell aus. Mit dem C5I-Campus haben wir einen Innovationsraum für hochsichere IT-Systeme geschaffen und bieten attraktivste Arbeitsplätze. Für fachliche Auskunft steht dir Marco Heinzen, Principal GIS Engineering C5I, gerne zur Verfügung: Tel. +41 79 568 14 96 | Mail marco.heinzen@ruag.ch
Deine Vorteile
Lohn und Nebenleistungen
Wir bieten dir ein leistungsorientiertes- und marktgerechtes Salär, 13 Monatslöhne sowie grosszügige Prämien- und Zulagen. Zusammen mit weiteren Nebenleistungen ergibt sich daraus ein attraktives Gesamtpaket.
Flexibles Arbeiten
Wir legen grossen Wert darauf, dass unsere Mitarbeitenden ein gesundes Gleichgewicht zwischen Beruf und Privatleben finden. Darum bieten wir unter anderem flexible Arbeitszeiten und Homeoffice-Optionen an.
Aufstiegs- und Weiterbildungsmöglichkeiten
Das Potenzial unserer Mitarbeitenden zu fördern und zu entfalten ist uns wichtig. Durch Schulungen, Kurse und Weiterbildungen unterstützen wir sie dabei, ihre Fähigkeiten kontinuierlich auszubauen um ihre beruflichen Ziele zu erreichen.
Wichtige Hinweise zum Bewerbungsprozess
* Vor Anstellungsbeginn wirst du darum gebeten, einen Straf- und Betreibungsregisterauszug vorzuweisen.
* Bewerbungen nehmen wir ausschliesslich über unser Stellenportal jobs.ruag.ch entgegen. E-Mail- wie auch Briefbewerbungen können wir leider nicht akzeptieren.
Zusätzliche offizielle RUAG-Karriereinformation, nicht Teil des obigen Stelleninserats:
RUAG führt bei allen neu eintretenden Mitarbeitenden eine Personensicherheitsprüfung durch.
output/RUAG_AI_Engineer_C5I/critique_ruag_ai_engineer_c5i.md
+200
View File
@@ -0,0 +1,200 @@
# Critique: RUAG C5I — AI Engineer C5I (w/m/d) — Round 2
**Date:** 2026-08-27 (Round 2, after Edit 1)
**Round 1:** same day, pre-edit. Evidence Fit 74 · Document Quality **85** · Channel Weak · 3 Tier 1 fixes.
**Documents:** `e2e_ruag_ai_engineer_c5i_resume.tex` (2 pp, **13 bullets**), `e2e_ruag_ai_engineer_c5i_cover_letter.tex` (1 p, **295 body words**, 5 short paragraphs + close)
**JD:** `JD_ruag_ai_engineer_c5i.txt` — verbatim live posting, retrieved 2026-08-27. **JD integrity: PASS.**
> Scored per `critique_framework.md`, which governs over `SKILL.md` §9.3/§9.4: separate Evidence Fit and Document Quality, verdicts rather than invented probabilities, no single overall score. Documents were re-read from disk, not scored from the Round 1 record.
---
## 1. Round 1 findings — resolution status
| # | Finding | Status |
|---|---|---|
| **T1-1** | Cover letter never connected to DevSecOps, C5I or the Army | **CLOSED.** New standalone paragraph carries `SW-5` and names the DevSecOps model and C5I |
| **T1-2** | `PP-3` missing — the only *current* Linux/ops evidence | **CLOSED.** New `Projekte` section; audited clean against PP-3's full forbidden list |
| **T1-3** | `governte Datenprodukte` is not a German word | **CLOSED.** Now `governance-konforme Datenprodukte` |
| **T2-2** | Headline dropped "AI" from a req titled AI Engineer | **CLOSED.** Now his canonical title, `Staff Data, Analytics & AI Engineer` |
| **T2-3** | Four honestly closeable JD terms absent | **CLOSED.** `Wissenstransfer`, `Datenaufbereitung`, `AI-/ML-Workloads`, `Containerisierung` |
| **T2-4** | IBM AI Engineering had no primary-source record | **CLOSED.** Certificate read directly: IBM via Coursera, 4 Jun 2020, verify `3ZBZFVAL6A34`. Recorded as entry #8 in `thiessen_certifications.md` |
| **T2-5** | CL paragraph 2 re-told resume bullets | **CLOSED.** Compressed; the working-style sentence survives |
| **T2-6** | Post-gap pivot conceded without evidence | **CLOSED.** Now anchored: *"wie zuvor bei der ML-Inferenz in Dresden"* |
| **T2-7** | Citizenship volunteered in the letter | **CLOSED by user decision** — cut. Resume header still carries it |
| **T3** | Bullets 9–10 both opened `Implementierte` | **CLOSED.** Bullet 9 now opens `Baute`; 0 consecutive identical openers |
| **T2-1(a)** | Vizrt title vs canonical `official_title` | **OPEN by user decision**, logged. Unchanged |
**Plus one defect Round 1 did not catch, found during the edit and fixed:** the resume **never loaded `babel`**, so a German document was hyphenated with English patterns (`Hal-bleiterfertigung`, `Tran-skription`, `automa-tisierte`). This was present at baseline and had survived a full visual QA, because nothing *looks* broken — the lines justify and the page is clean. Fixed with `\usepackage[ngerman]{babel}` plus a `\hyphenation{Trans-krip-ti-on}` exception, which also cleared an overfull box the longer headline had introduced.
---
## 2. Fit Verdict and Hard Gates — unchanged, and unchangeable by editing
| Gate | Result |
|---|---|
| Title-defining capability Direct? | **FAIL.** R1 (build/operate an internal AI-/LLM platform incl. compute cluster) and R2 (in-house LLM solutions, connectors, retrieval/document solutions) remain **Adjacent** |
| Minimum qualification a Gap? | **PASS.** None of the eight `Das bringst du mit` items is a Gap |
| Level requires absent scope? | **PASS** |
| Location / authorization | **PASS.** Thun role, Thun resident, German native, EU citizen with Swiss B permit |
| Clearance (PSP) | **UNRESOLVED** |
| Compensation / level | **UNRESOLVED** |
No document change touches any of this. The user overrode this gate on 2026-08-27 with all of it in view; recorded, not re-litigated.
---
## 3. Evidence Fit — 74/100 (unchanged)
| Dimension | Weight | Score | Reasoning |
|---|---:|---:|---|
| Required qualifications | 35 | 29 | Degree, years, Python, Linux/Docker/Kubernetes, German/English Direct. ML/AI-framework half of Q3 Adjacent |
| Core responsibilities | 25 | 17 | R3/R5/R6 Direct, R4 Direct-to-Adjacent; R1 and R2 Adjacent |
| Level and ownership | 15 | 11 | Staff plus Component/Application Owner exceeds "Berufserfahrene"; may be lateral or down-level |
| Recency and depth | 10 | 8 | Python/Kubernetes/data current; production-ML integration ended 2022; LLM evidence configuration-level |
| Domain/tool transfer | 10 | 6 | Compute cluster, GPU, RAG and search are not substitutable by wording |
| Practical constraints | 5 | 3 | Location/language/authorization excellent; PSP and compensation unresolved |
| **Total** | **100** | **74** | Numerically top-of-Adjacent; strategically **Stretch** |
**`PP-3` does not move this score.** It strengthens the *document's* evidence for the operating half of the role, but it is a personal project and cannot convert an Adjacent title-capability into a Direct one. Treating it as fit-moving would be exactly the "bridge valued as highly as direct evidence" error §5 warns against.
---
## 4. Document Quality — 93/100 (was 85)
| Dimension | Weight | R1 | R2 | Reasoning for the change |
|---|---:|---:|---:|---|
| Truth and provenance | 25 | 23 | **24** | IBM AI Engineering now has a primary source, closing the one line a canonical preflight could not confirm. New `PP-3` and `SW-5` content audited clean against their full forbidden lists. The headline is now *more* accurate than before — his real title, not a positioning label he never held. Remaining deduction: the Vizrt title, user-authorized and logged, is still the one line that will not match a Zeugnis |
| Information hierarchy | 20 | 17 | **19** | Headline now carries the req's own word above the fold. `Projekte` is cleanly placed and parses correctly. Remaining: the officer career, an explicit JD advantage, still sits on page 2 |
| Bullet evidence and impact | 20 | 16 | **16** | **Unchanged and still the weakest dimension.** 12 of 13 bullets are activity descriptions with no outcome clause; only `BS-1` carries a result. This is the honest consequence of `metrics: unverified` across the KB, not a writing failure — but it is real, and it is what a hiring manager notices |
| Relevance and terminology | 15 | 12 | **15** | Both Round 1 deductions closed. **22/22 claimable JD terms now present**, and all six gap terms (RAG, Retrieval, Compute Cluster, AI-/LLM-Plattform, Connectors, Dokumentenlösungen) remain correctly absent. `MVP/Pilotlösung` and `R&D` stay uncovered because the evidence supports "Proof of Concept" and nothing more — declining to stretch is correct behaviour, not a deduction |
| Skills evidence | 10 | 9 | **9** | All tokens still traced. `Datenaufbereitung` and `AI-/ML-Workloads` are claim-backed. TensorFlow/Keras now rests on a verified certificate. Unchanged deduction: `MLOps` in the headline is an umbrella term with no canonical `skills[]` entry, claim-backed only via BS-1 |
| Mechanics/readability | 10 | 8 | **10** | `governte` fixed. German hyphenation fixed. Cadence: list-terminated bullets **50% → 31%**, consecutive identical openers **1 → 0**. 0 boxes, 0 warnings, correct parse, consistent Swiss orthography |
| **Total** | **100** | **85** | **93** | |
**Truth and provenance 24/25 (9.6/10)** — well clear of the 8/10 automatic-failure floor.
The 8-point rise is almost entirely the closure of named Round 1 deductions plus a defect Round 1 missed. The one dimension that did **not** move is bullet impact, and it is the one that would most change a hiring manager's read — see Tier 2.
---
## 5. Channel Strength — **Weak** (unchanged)
Cold portal application; `jobs.ruag.ch` refuses email and post. **Marco Heinzen's direct mobile (+41 79 568 14 96) is published in the posting and still unused.** Nothing in Edit 1 changed this, and no further document work can. This remains the single highest-value action on the application: one call upgrades the channel to Moderate *and* answers whether C5I is hiring someone who has built an AI platform or someone who can operate one and grow into building it — the question the whole application turns on.
---
## 6. Reader Sequence — Verdicts
**Parser / ATS (Umantis). PASS.** `pdftotext -layout` extracts employer, title, dates, location and section order correctly on both pages; the new `Projekte` block parses cleanly; footers read 1/2 and 2/2. No tables, text boxes or multi-column body.
**Recruiter glance (10 s) — Frank Haugwitz. FORWARD.** *The Round 1 friction point is gone.* He is filling a req titled **AI Engineer**, and the headline now reads `Staff Data, Analytics & AI Engineer` — the candidate's real title, containing the word, above the fold, alongside Thun, citizenship, permit and native German. Every box he owns is answered before he scrolls.
**HR / dossier completeness (30 s). COMPLETE.** Two pages, no photo, clean chronology, education, languages, work authorization, `Militärischer Werdegang`, and now a `Projekte` section — which reads as a fuller local dossier, appropriate for a traditional Swiss employer. The Nov 2014 – Mai 2015 gap remains intentional and unremarkable.
**Hiring manager (2 min) — Marco Heinzen. Still MAYBE, but a stronger MAYBE.** The underlying split is unchanged: if his brief is "build our RAG stack", this is a no; if it is "run the platform and the pipelines while we grow the AI layer", it is a strong yes. What changed is that the letter now argues the second case more completely — it names his DevSecOps practice against C5I's stated operating model, and the pivot after the gap admission is anchored in something he has actually done rather than in an intention. The resume now also shows *current* hands-on infrastructure, not only work that stopped in 2022.
**Technical reviewer (10 min). CREDIBLE.** No claim collapses. Likely probes, in order: (1) "You configured the assistants — who built the platform underneath?" (2) "Compute cluster — what is the largest thing you have operated, and have you scheduled GPU workloads?" (3) **new:** "Your private server — what have you actually hardened, and how do you patch it?" This is a probe the document now *invites*, and it is a good one to invite. (4) The 2008–2014 / 2009–2013 officer-track overlap. (5) Ansible vs Pull-GitOps — the word is still correctly absent and the substance is answerable.
---
## 7. Claim Audit — new and changed content only
| Claim | Canonical | Safe? | Note |
|---|---|---|---|
| Headline `Staff Data, Analytics & AI Engineer` | `claims.json` SWISSCOM `title_history` / current title | ✓ | Now his **actual** title. More accurate than the Round 1 headline |
| Profil opener, same title | as above | ✓ | Consistent with headline and body |
| `governance-konforme Datenprodukte innerhalb des unternehmensweiten AWS Data Mesh` | SW-7 | ✓ | Object still scoped; mesh still not claimed |
| **`Projekte`: private Debian server** | **PP-3** | ✓ | Verbs `Betreibt`/`administriert`/`verantwortet Systemhärtung` all in `allowed_verbs`. "selbst gehostet" stated; section reads `Eigenbetrieb`. **No** enterprise/production/multi-user/business-critical wording, **no** uptime/availability/SLA/user-count/scale figure, **no** SRE or sysadmin implication, **no** datacenter claim. Audited against all five forbidden items |
| `koordinierte Hersteller, technische Dokumentation und Wissenstransfer an die Nutzer` | BS-3 | ✓ | All four BS-3 elements retained (SLOs, vendors, documentation, training) in natural German |
| `Baute einen ELK/Kafka-Proof-of-Concept` | BS-4 | ✓ | `built` is an allowed verb; PoC status still explicit |
| Skills `Datenaufbereitung` | SW-1 / SW-2 / BS-2 | ✓ | Real work, previously unnamed |
| Skills `Containerisierung und Deployment von AI-/ML-Workloads` | BS-1 / BS-6 | ✓ | No new claim; renames existing evidence in the JD's vocabulary |
| Cert `IBM AI Engineering mit TensorFlow/Keras im Weiterbildungskontext` | `thiessen_certifications.md` #8 | ✓ | **Now primary-source verified.** Still correctly confined to certification context |
| **CL:** `Für 2025/2026 bin ich Security Champion meines Teams, mit 100 Stunden Training…` | SW-5 | ✓ | Single year only. Team role, not an award. Does **not** claim DevSecOps compliance ownership |
| **CL:** `Dass C5I konsequent nach dem DevSecOps-Modell arbeitet, entspricht damit meiner heutigen Praxis` | SW-5 + SW-3 | ✓ | A soft alignment claim, supported by the Security Champion role, 100 h training and GitLab CI/CD automation. See Tier 2 — expect it to be probed |
| **CL:** `die Implementierungsseite erarbeite ich mir von dort aus, wie zuvor bei der ML-Inferenz in Dresden` | BS-1 | ✓ | Back-reference to real work; introduces no new claim |
| Citizenship sentence removed from CL | — | ✓ | Resume header still carries `Deutscher Staatsbürger` / `Schweizer Aufenthaltsbewilligung B`, so nothing is concealed |
**No Tier 1 truth finding.** Full sweep across both `.tex` sources and both extracted PDFs: **0 hits** for Capgemini, LangChain/LangGraph/LlamaIndex, Terraform, Azure, GCP, GitOps, TypeScript, FastAPI, Flask, Django, every `global_forbidden_output_patterns` entry, and the standalone acronym `RAG` (word-boundary count 0 — the only `rag` matches are inside *Fragen* and *Abfragen*). **0 em dashes.**
---
## 8. Tiered Improvements
### Tier 1 — none.
No truth, fit or relevance defect remains that is fixable by editing. This is the honest result of Round 1's fixes, not a courtesy.
### Tier 2
**T2-A · Bullet impact is now the document's weakest dimension, and it is a KB problem, not a writing problem.**
12 of 13 bullets describe activity with no outcome. `BS-1` is the only bullet that says what changed. Every other claim in `claims.json` carries `metrics: unverified`, and inventing figures is barred — so the resume is correctly written but structurally flat where a hiring manager looks for consequence. **The fix is upstream:** if any verified outcome exists for `SW-2` (incident or SLA record), `SW-3` (deployment frequency, service count), `SW-7` (data products delivered, sources onboarded) or `BS-3` (applications owned), recording it in `claims.json` would raise this dimension across every future package, not just this one. Do not invent anything to close it here.
**T2-B · The officer career still sits on page 2.**
The JD names it explicitly as an advantage, and it is the one thing most competing software candidates cannot offer. It is mitigated by the profile sentence, and moving the section would break reverse chronology. Structural; accepted rather than fixed.
**T2-C · The headline now duplicates the Swisscom title line verbatim.**
Both read `Staff Data, Analytics & AI Engineer`. Accurate and defensible — it is his title — but the repetition is visible within the first fifteen lines. Optional: differentiate the headline's tail (it currently carries `Python · Kubernetes · MLOps`) or leave it. Cosmetic only.
**T2-D · Prepare the DevSecOps sentence for a follow-up question.**
*"…entspricht damit meiner heutigen Praxis"* is a soft alignment claim. It is supported — Security Champion, 100 h training, GitLab CI/CD — but a security-conscious reader may ask what DevSecOps actually looks like in his current team, and the honest answer is that he is the team's security point of contact, not the owner of a DevSecOps pipeline. Worth having ready; not worth rewording.
### Tier 3
- **Bosch bullet lengths cluster** at 19/22/17/18 words — a 5-word spread, just above §7a's flagged threshold of ≤4. The validator passes it. Diagnostic only; do not pad or trim to move it.
- **`\today`** renders "27. August 2026". If the built PDF is submitted later, recompile so the date refreshes.
- **`MLOps`** in the headline remains an umbrella term with no canonical `skills[]` entry. Claim-backed via BS-1; keep.
---
## 9. Cover Letter Critique
**6A — Anti-patterns. PASS.** No generic opener, no company-history paraphrase, no adjective stack, no defensive gap list. Opens on the candidate-role connection in the first clause. 0 em dashes. The gap sentence — *"Eine AI-Plattform habe ich nicht aufgebaut"* — remains one sentence, immediately followed by what he does bring, and is now followed by evidence rather than intention.
**6B — Tailoring. PASS (was Partial).** The Round 1 deduction is closed. The letter now engages the posting's own `Über den Bereich` framing by name — **C5I** and the **DevSecOps model** — and does so through evidence that was already on the resume rather than through company-research filler. The role thesis in paragraph 1 still derives directly from the JD's first two bullets.
**6C — Context-specific. PASS.** Defence motivation correctly bounded: *"im Umfeld von Landesverteidigung und gesellschaftlicher Verantwortung"*, never rejoining the armed forces. *"ich wohne in Thun, also am Standort selbst"* remains the strongest practical line in the package.
**6D — ATS.** Not a factor; the letter supplements the CV in a portal upload.
**6E — Structural. PASS.** 295 body words (target 250–300). Six blocks with good variation — 50 / 52 / 77 / 35 / 63 / 18 words — and no paragraph dominating the page after the split. One page, 0 boxes. Swiss orthography consistent (*Grüsse*, *einschliesslich*). Salutation correctly without a comma. Addressed to the named Ansprechperson at the verbatim-verified legal entity and address. `cl_reference.md`'s "two or three short paragraphs" is the International-Tech default; line 49 explicitly permits an employer-appropriate motivation-statement format, and five short German paragraphs serve this reader better than three long ones.
**6F — Package cohesion. PASS.** Every letter claim maps to a resume bullet and a canonical ID (BS-1, BS-6, SW-3, SW-2, SW-8, **SW-5**, BW-1). No new tool, metric, customer or ownership claim. The LLM boundary is stated identically in both documents. Contact block and location consistent. Citizenship now appears in exactly one place — the resume header — which is the right number.
**Would deleting the letter improve the application? No.** It is now doing more work than in Round 1: it answers the question the resume structurally cannot — what this candidate thinks he is applying for, given what he has not done — and it connects his current security practice to the employer's stated operating model.
---
## 10. Mechanical Verification — re-run on the edited files
| Check | Resume | Cover letter |
|---|---|---|
| `validate_resume_system.py --document` | **PASS, 0 warnings** | **PASS, 0 warnings** |
| Canonical system validator | **PASS, 0 warnings** | — |
| `claims.json` parses after the PyTorch patch | **valid JSON** | — |
| MiKTeX, two passes | **PASS** | **PASS** |
| Page count | **2** (exact) | **1** (exact) |
| Overfull / Underfull boxes | **0** | **0** |
| LaTeX warnings | **0** | **0** |
| `pdftotext -layout` order | employer / title / date / location / sections correct incl. new `Projekte`; footers 1/2, 2/2 | correct |
| German hyphenation | **ngerman** — `Halb-leiterfertigung`, `Trans-krip-tion`, `automati-sierte` | ngerman (was already loaded) |
| Visual QA at 130 dpi | both pages re-rendered and inspected after every edit | re-rendered and inspected after the split |
| Submission copy refreshed + MD5 match | **yes** | **yes** |
| PDF newer than `.tex` | **yes** — no unbuilt edits pending | **yes** |
**Content checks.** Email `dennis@thiessen.io` ✓. `Thun, Schweiz` on both; Swisscom entry keeps `Bern, Schweiz` ✓. Capgemini 0 ✓. All forbidden skills 0 ✓. All global-forbidden patterns 0 ✓. French/Italian absent ✓. No DOB, marital status, gender, children or photo ✓. No LOC or test counts ✓. Certifications appear exactly once ✓.
**Structural.** 13 bullets (11–14 guide ✓). Six skills lines (4–6 guide, sixth is Zertifizierungen ✓). Profile 4 rendered lines, one over the 2–3 guide but carrying three distinct positioning claims; not flagged. Reverse chronology intact. `Militärischer Werdegang` and `Projekte` correctly separated from `Berufserfahrung`.
---
## 11. Summary
**Evidence Fit 74/100 · Stretch · title-capability gate FAIL · PSP and compensation UNRESOLVED — all unchanged.**
**Document Quality 93/100 (was 85) · truth and provenance 24/25 · no Tier 1 findings of any kind.**
**Channel Weak — unchanged, and the published direct line is still unused.**
The documents are now as good as this evidence base allows. Every Round 1 finding is closed except the Vizrt title, which is a conscious user decision, and bullet impact, which is a `claims.json` limitation rather than a writing one.
Per `critique_framework.md`, a high Document Quality score does not make this package submit-ready, and 93 does not mean 93% likely to interview. The gate still reads FAIL. What decides this application is not the documents — it is whether C5I wants the platform's AI layer or its operating layer, and that is answered by a phone call, not by another edit.
output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_cover_letter.tex
+69
View File
@@ -0,0 +1,69 @@
% RUAG C5I: AI Engineer C5I (w/m/d). German-language Swiss/DACH letter.
% Every claim traces to a resume bullet and a canonical ID.
% Boundaries enforced here:
% - SW-8 stays configuration-level. No RAG, retrieval, orchestration, evaluation or fine-tuning.
% - No AI/LLM-platform ownership and no compute/GPU-cluster operation is implied anywhere.
% - BS-1 does not claim model training.
% - BW-1: no clearance, no command scope, no technical military work.
% - RUAG is a defence contractor, not the armed forces. The letter says "im Umfeld von
% Landesverteidigung", never "wieder Teil der Armee".
% - The LLM gap is named once, briefly and without apology. The JD invites applicants who do
% not meet every requirement, so an honest bridge reads as candour, not as a defensive list.
% - SW-5 (Security Champion 2025/2026, 100 h training) answers the JD's stated DevSecOps model.
% Team role only, never an award, never DevSecOps compliance ownership.
% - Citizenship is deliberately NOT restated here (user decision 2026-08-27); the resume header
% already carries "Deutscher Staatsbuerger" and "Schweizer Aufenthaltsbewilligung B".
% Contact location is Thun, matching the resume and the Thun-area rule in config.md.
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[ngerman]{babel}
\usepackage{lmodern}
\usepackage[a4paper,left=0.85in,right=0.85in,top=0.7in,bottom=0.7in]{geometry}
\usepackage{parskip}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\begin{document}
{\Large\bfseries Dennis Thiessen, M.Eng.}\par
Thun, Schweiz $\vert$
\href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$
\href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}
\vspace{1.5em}
RUAG MRO Holding AG, Business Area C5I\\
Herr Frank Haugwitz\\
Uttigenstrasse 36, 3600 Thun
\vspace{1em}
\today
\vspace{1em}
\textbf{Bewerbung als AI Engineer C5I (w/m/d)}
\vspace{0.8em}
Sehr geehrter Herr Haugwitz
Ihre Ausschreibung verbindet zwei Aufgaben, die in meiner Arbeit ohnehin zusammengehören: eine interne Plattform für AI-Workloads zu betreiben und die Datenpipelines und Schnittstellen zu bauen, von denen diese Workloads leben. An dieser Schnittstelle arbeite ich seit Jahren, zuerst in der Halbleiterfertigung bei Bosch, heute bei Swisscom auf Kubernetes und AWS.
Bei Bosch in Dresden integrierte ich ML-Inferenz als Docker-Container mit Kubernetes und Ansible in die kontinuierlich laufende Fertigung; die Klassifikation von Defektbildern lief danach vollautomatisiert. Die Linux-Server darunter administrierte ich selbst und automatisierte ihre Konfiguration mit Ansible, inklusive eigener Erweiterungen. In einer Fertigung ohne Stillstandstoleranz zeigt sich früh, ob eine Automatisierung trägt.
Heute entwickle und betreibe ich bei Swisscom Python-Datenanwendungen auf Kubernetes und verantworte als Component Owner geschäftskritische ETL-Strecken von Oracle und Kafka ins Data Warehouse, einschliesslich Rufbereitschaft. Meine LLM-Erfahrung möchte ich klar einordnen: Ich konfiguriere domänengestützte Assistenten in einer Swisscom-eigenen Oberfläche und arbeite über LiteLLM mit Modell-APIs. Eine AI-Plattform habe ich nicht aufgebaut. Was ich mitbringe, ist die Betriebs- und Plattformseite dieser Rolle; die Implementierungsseite erarbeite ich mir von dort aus, wie zuvor bei der ML-Inferenz in Dresden.
Sicherheit ist dabei kein Anhang: Für 2025/2026 bin ich Security Champion meines Teams, mit 100 Stunden Training zu Cloud Security, DevSecOps und Security by Design. Dass C5I konsequent nach dem DevSecOps-Modell arbeitet, entspricht damit meiner heutigen Praxis.
Die Aufgabe hat für mich auch persönlich Gewicht. Ich diente sechs Jahre in der Bundeswehr, absolvierte den Offizieranwärterlehrgang und die Offizierschule und schied als Leutnant aus. Nach über zehn Jahren in der Industrie wieder im Umfeld von Landesverteidigung und gesellschaftlicher Verantwortung zu arbeiten, ist ein wesentlicher Grund für diese Bewerbung. Deutsch ist meine Muttersprache, und ich wohne in Thun, also am Standort selbst.
Über ein Gespräch würde ich mich freuen, besonders darüber, wie Plattform und erste Inhouse-Lösungen bei Ihnen zusammenspielen sollen.
Freundliche Grüsse
\vspace{1.5em}
Dennis Thiessen
\end{document}
output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_resume.tex
+92
View File
@@ -0,0 +1,92 @@
\documentclass{resume}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[ngerman]{babel}
\usepackage{lmodern}
\usepackage[a4paper,left=0.65in,right=0.65in,top=0.55in,bottom=0.55in]{geometry}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\usepackage[version=4]{mhchem}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[R]{\small \thepage/\pageref{LastPage}}
\hyphenation{Trans-krip-ti-on}
\name{Dennis Thiessen, M.Eng.}
\headline{Staff Data, Analytics \& AI Engineer $\vert$ Python · Kubernetes · MLOps}
\contactline{Thun, Schweiz $\vert$ \href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$ \href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}\\Deutscher Staatsbürger $\vert$ Schweizer Aufenthaltsbewilligung B\\Deutsch (Muttersprache) $\vert$ Englisch (fliessend)}
\begin{document}
\begin{rSection}{Profil}
Staff Data, Analytics \& AI Engineer mit über zehn Jahren Erfahrung in produktiven Daten- und Softwaresystemen. Verbindet Python/Kubernetes und Linux-/Ansible-Automatisierung mit produktiver ML-Inferenz und der Konfiguration domänengestützter LLM-Assistenten; sechs Jahre Bundeswehr mit abgeschlossener Offizierausbildung ergänzen das Profil.
\end{rSection}
\begin{rSection}{Kenntnisse}
\skillline{Programmierung \& Schnittstellen}{Python, SQL, REST APIs; LiteLLM in aktueller API-Nutzung}
\skillline{AI/ML Engineering}{Containerisierung und Deployment von AI-/ML-Workloads, domänengestützte LLM-Assistenten, NLP/Spracherkennung im Forschungsprojekt}
\skillline{Plattform \& Automation}{Linux, Docker, Kubernetes, Ansible, GitLab CI/CD, Jenkins}
\skillline{Data Engineering}{Kafka, Airflow, PySpark, AWS (S3, Glue, Athena/Iceberg, Redshift), Datenpipelines, Datenaufbereitung}
\skillline{IaC \& Betrieb}{CloudFormation, ELK, Grafana, Prometheus, On-Call-Betrieb, DevSecOps}
\skillline{Zertifizierungen}{AWS Solutions Architect -- Associate, gültig bis Sep. 2027; Data Engineering with AWS, 2026; IBM AI Engineering mit TensorFlow/Keras im Weiterbildungskontext; iSAQB CPSA Foundation}
\end{rSection}
\begin{rSection}{Berufserfahrung}
\begin{rSubsection}{Swisscom (Schweiz) AG}{Okt. 2023 -- heute}{Staff Data, Analytics \& AI Engineer (Engineer IV; Beförderung Apr. 2025)}{Bern, Schweiz}
\item Konfiguriert domänengestützte LLM-Assistenten in einer Swisscom-eigenen Weboberfläche: Auswahl verfügbarer Modelle und kuratierte Wissensgrundlagen für Fragen, Migrationsunterstützung und Datenmapping.
\item Entwickelt, deployt und betreibt Python-Datenanwendungen auf Kubernetes; automatisiert Build, Tests und Deployment mit GitLab CI/CD.
\item Modelliert und implementiert governance-konforme Datenprodukte innerhalb des unternehmensweiten AWS Data Mesh von Swisscom. Onboardet Quellsysteme und dokumentiert Metadaten sowie Lineage in Atlassian Compass.
\item Verantwortet als Component Owner geschäftskritische Fulfillment-ETL-Pipelines von Oracle und Kafka in das Teradata DWH, einschliesslich Data Governance, 2nd/3rd-Level-Support, Rufbereitschaft und SLA-Einhaltung.
\item Übernimmt 2025/2026 die Teamrolle als Security Champion; absolvierte 100 Stunden Training zu Cloud Security, DevSecOps, Security by Design und Risikomanagement mit abschliessender Prüfung.
\end{rSubsection}
\end{rSection}
\newpage
\begin{rSection}{Berufserfahrung (Fortsetzung)}
\begin{rSubsection}{Robert Bosch Semiconductor Manufacturing Dresden GmbH}{Feb. 2020 -- Dez. 2022}{Senior Data Engineer, Data Analysis (Beförderung Jan. 2021)}{Dresden, Deutschland}
\item Integrierte ML-Inferenz als Docker-Container mit Kubernetes und Ansible in die kontinuierliche Halbleiterfertigung. Ermöglichte damit eine vollautomatisierte Klassifikation von Defektbildern.
\item Administrierte Linux-Server für Analyse- und ML-Workloads und automatisierte deren Konfiguration mit Ansible. Entwickelte ein Plugin, das Zugangsdaten lokal zwischenspeicherte und Keystore-Abfragen reduzierte.
\item Verantwortete als Application Owner Analyseanwendungen und vorgelagerte Pipelines; definierte SLOs und koordinierte Hersteller, technische Dokumentation und Wissenstransfer an die Nutzer.
\item Baute einen ELK/Kafka-Proof-of-Concept für Anomalieerkennung und ergänzte das Monitoring um Grafana, Prometheus und Loki.
\end{rSubsection}
\begin{rSubsection}{Fraunhofer-Center für Maritime Logistik und Dienstleistungen CML}{Sep. 2018 -- Okt. 2019}{Wissenschaftlicher Mitarbeiter / Research Software Engineer}{Hamburg, Deutschland}
\item Implementierte im ARTUS-Forschungsteam ML- und NLP-Komponenten für die automatische Transkription von Kommunikation bei Seenotrettungseinsätzen.
\end{rSubsection}
\begin{rSubsection}{Vizrt}{Juli 2017 -- Mai 2018}{DevOps Engineer}{Bergen, Norwegen}
\item Entwickelte Python-Backend-Komponenten für verteiltes Video-Transcoding und integrierte automatisierte Audio-/Video-Tests als Quality Gates in die CI/CD-Pipeline.
\end{rSubsection}
\compactentry{Generali Deutschland Informatik Services GmbH}{Mai 2015 -- Juni 2017}
\textit{IT Consultant}\hfill\textit{Hamburg, Deutschland}\par
\end{rSection}
\begin{rSection}{Militärischer Werdegang}
\begin{rSubsection}{Bundeswehr}{Juli 2008 -- Nov. 2014}{Offizieranwärter / Offizier}{Deutschland}
\item Absolvierte Offizieranwärterlehrgang und Offizierschule während sechsjähriger Dienstzeit; schied im Dienstgrad Leutnant aus der Bundeswehr aus.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Projekte}
\begin{rSubsection}{Privater Debian-Server}{seit rund zehn Jahren}{Eigenbetrieb}{}
\item Betreibt und administriert einen selbst gehosteten Debian-Server mit nginx, Mailserver, Nextcloud, VPN, Docker-Diensten und Bitwarden; verantwortet Systemhärtung und Security-Konfiguration selbst.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Ausbildung}
\compactentry{M.Eng. Computer Aided Engineering}{Apr. 2012 -- Okt. 2013}
Universität der Bundeswehr München, Vertiefung Software Design \& Engineering. Masterarbeit an der Tongji University, Shanghai: \textit{Development of a Web-Based Remote Fault Diagnosis System}; Note 1,0.
\vspace{0.25em}
\compactentry{B.Eng. Information and Telecommunication Technologies}{Okt. 2009 -- Okt. 2012}
Universität der Bundeswehr München.
\end{rSection}
\end{document}
output/RUAG_AI_Engineer_C5I/resume.cls
+58
View File
@@ -0,0 +1,58 @@
\ProvidesClass{resume}[2026/07/27 Evidence-first professional resume]
\LoadClass[11pt,a4paper]{article}
\RequirePackage{enumitem}
\RequirePackage{ifthen}
\RequirePackage{lastpage}
\RequirePackage{parskip}
\RequirePackage{needspace}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\def\@resumename{}
\def\@resumeheadline{}
\def\@resumecontact{}
\newcommand{\name}[1]{\def\@resumename{#1}}
\newcommand{\headline}[1]{\def\@resumeheadline{#1}}
\newcommand{\contactline}[1]{\def\@resumecontact{#1}}
\newcommand{\printresumeheader}{%
{\LARGE\bfseries \@resumename}\par
\vspace{0.15em}
\ifthenelse{\equal{\@resumeheadline}{}}{}{\@resumeheadline\par}
\vspace{0.1em}
{\small \@resumecontact}\par
\vspace{0.45em}
}
\AtBeginDocument{\printresumeheader}
\newenvironment{rSection}[1]{%
\vspace{0.35em}
{\bfseries #1}\par
\vspace{0.1em}\hrule\vspace{0.35em}
}{%
\vspace{0.2em}
}
\newenvironment{rSubsection}[4]{%
\Needspace{6\baselineskip}%
{\bfseries #1}\hfill{\small #2}\par
{\itshape #3}\hfill{\itshape #4}\par
\vspace{0.1em}
\begin{itemize}[leftmargin=1.25em,label=\textbullet,itemsep=0.18em,topsep=0.15em,parsep=0pt,partopsep=0pt]
}{%
\end{itemize}
\vspace{0.25em}
}
\newcommand{\skillline}[2]{%
\textbf{#1:} #2\par
\vspace{0.08em}
}
\newcommand{\compactentry}[2]{%
\textbf{#1}\hfill #2\par
}
output/RUAG_AI_Engineer_C5I/session_ruag_ai_engineer_c5i.md
+474
View File
@@ -0,0 +1,474 @@
# Session: RUAG C5I - AI Engineer C5I (w/m/d)
## JD Integrity
- **File/source:** `JD_ruag_ai_engineer_c5i.txt`; verbatim visible body from `https://jobs.ruag.ch/offene-stellen/ai-engineer-c5i/4afaa9aa-74c1-4b4d-9232-cf65651ce8ec`.
- **Retrieval method and date:** direct web retrieval of the live RUAG posting and apply form, 2026-08-27.
- **Verbatim posting:** YES.
- **Posting status:** LIVE 2026-08-27; application form reachable. Internal application ID 18027, publish ID 10075372.
- **Official portal listing:** Berufserfahrene, Thun, Switzerland, 80-100%.
## Application Decision
- **Audience profile:** Swiss/DACH. German-language posting, Swiss defence employer, local Thun role.
- **Evidence Fit:** **74/100**.
- **Fit class:** **Stretch under `application_strategy.md`**, although the numerical score is at the top of Adjacent. The title-defining AI-/LLM-platform build and in-house LLM implementation are only Adjacent.
- **Hard gate:** **FAIL under the strict `critique_framework.md` title-capability rule.** Dennis has not built or operated an AI-/LLM platform or implemented a production retrieval/document solution. His evidence is real but one layer adjacent: Kubernetes/data-platform operation, professional Linux/Ansible, containerized ML inference integration, LiteLLM/API exposure and configuration of grounded LLM assistants.
- **Channel Strength:** Weak today; upgradeable to Moderate only after an actual conversation.
- **Channel plan:** contact Frank Haugwitz (`recruiting-c5i@ruag.ch`) for compensation/PSP screening and Marco Heinzen (`marco.heinzen@ruag.ch`, +41 79 568 14 96) for technical scope, especially whether the first six months centre on platform engineering or direct RAG/LLM implementation.
- **Cohort slot:** evidence-first-2026-01 currently 5/10: Core 3/7, Adjacent 2/2, Stretch 0/1. Proceeding would consume the sole Stretch slot.
- **Decision:** **PROCEED by explicit user override, 2026-08-27.** The user consciously accepts the title-capability gap, unresolved compensation/PSP questions and use of the cohort's sole Stretch slot. The Evidence Fit, Stretch classification and strict hard-gate result remain unchanged.
## Role and Practical Context
| Field | Current fact |
|---|---|
| Employer | RUAG, Business Area C5I |
| Role | AI Engineer C5I (w/m/d) |
| Location | C5I, Uttigenstrasse 36, 3600 Thun |
| Pensum | 80-100% |
| Target group | Berufserfahrene |
| Technical contact | Marco Heinzen, Principal GIS Engineering C5I, +41 79 568 14 96, `marco.heinzen@ruag.ch` |
| Recruiting contact | Frank Haugwitz, Director Sourcing C5I-Campus, `recruiting-c5i@ruag.ch` |
| Application documents | CV / complete dossier required; motivation letter and certificates optional |
| Work model | Flexible hours and home-office options stated; no cadence promised. Commute is immaterial because the role is in Thun |
| Security process | RUAG states that all new employees undergo a Personensicherheitsprüfung; the JD also requests criminal- and debt-register extracts before employment |
| Compensation | No band. RUAG states market-based pay, 13 monthly salaries and allowances. Small self-reported samples indicate a meaningful risk of a cut versus current approximately CHF 140k; exact role band must be asked |
### PSP / citizenship read
- The posting contains **no Swiss-citizenship requirement**.
- RUAG's official careers page says all new employees undergo a PSP.
- SEPOS states that people resident abroad can in principle undergo a PSP and that third-party/company employees are checked for security-sensitive work. This does not guarantee a positive result or this role's project eligibility.
- German citizenship plus Swiss B work authorization is therefore **not an evidenced automatic disqualifier**, but the role-specific PSP/project-access position remains a practical question for RUAG.
### Compensation read
- Official posting: no salary band; only market-based pay, 13 salaries and benefits.
- Kununu RUAG samples are small and non-authoritative: DevOps Engineer approximately CHF 96.6k (n=2), Data Scientist approximately CHF 106.2k (n=2), Software Developer approximately CHF 110.5k average with CHF 78.3k-138.4k range (n=6).
- These numbers cannot price this vacancy, but they make a material pay cut versus current compensation plausible. The Thun-local exception permits near-current pay, not a large reduction.
## Requirements
| # | Requirement | Req/Pref | Class | Canonical evidence | Gate? |
|---|---|---|---|---|---|
| Q1 | BSc/MSc in Informatics, Computer Science, Data Science or comparable | Required | **Direct** | EDU-MENG (M.Eng. Computer Aided Engineering, Software Design & Engineering); EDU-BENG | - |
| Q2 | Several years in ML Engineering, Data Engineering, Software Engineering or DevOps | Required, disjunctive | **Direct** | SWISSCOM employment + SW-1/SW-2/SW-3; BOSCH + BS-1/BS-2/BS-6; FRAUNHOFER; VIZRT | - |
| Q3 | Very strong Python and experience with modern ML/AI frameworks | Required | **Split: Python Direct, frameworks Adjacent** | Python production-current; FC-2 ML/NLP contribution; BS-1 ML-inference integration; TensorFlow/Keras only certification-context evidence | ⚠ |
| Q4 | Linux, Docker, Kubernetes or comparable platforms | Required, disjunctive | **Direct** | BS-6, PP-3, BS-1, SW-3 | - |
| Q5 | LLM, RAG, APIs, data pipelines or search knowledge | Preferred, disjunctive | **Direct overall / partial** | SW-8 LLM configuration and LiteLLM API exposure; SW-1/SW-2/SW-3 data pipelines. RAG and search implementation remain Gaps | - |
| Q6 | Structured, independent, pragmatic execution | Required behaviour | **Adjacent** | SW-2 Component Owner, BS-3 Application Owner, FC-1 independent CI/CD setup; do not turn these into unsupported personality claims | - |
| Q7 | Very good German; English preferred | Required / preferred | **Direct** | German native, English fluent | - |
| Q8 | Higher NCO or officer career | Preferred | **Direct and differentiating** | BW-1: six years Bundeswehr, officer training/school, left as Second Lieutenant | - |
| C1 | Swiss work authorization | Practical | **Direct** | German citizen, Swiss B permit, no sponsorship required | - |
| C2 | PSP and project access | Practical | **Constraint** | PSP applies to all RUAG entrants; no nationality gate stated; role-specific outcome unknown | ⚠ |
| C3 | Compensation acceptable | Practical | **Constraint** | No published band; external samples suggest downside risk | ⚠ |
## Core Responsibilities
| # | Responsibility | Class | Canonical evidence | Gap boundary |
|---|---|---|---|---|
| R1 | Build, evolve and operate internal AI/LLM platform including compute cluster | **Adjacent, title-defining** | SW-3 Kubernetes/Python operations; BS-6 professional Linux/Ansible; BS-1 ML inference | No AI/LLM platform ownership, compute-cluster or GPU-cluster operation |
| R2 | Implement in-house LLM solutions, connectors, tools, retrieval/document solutions | **Adjacent-to-Gap, title-defining** | SW-8 configured grounded assistants; LiteLLM API hands-on | No production LLM implementation, RAG/retrieval implementation, orchestration or formal evaluation |
| R3 | Develop data pipelines, preprocessing and internal interfaces | **Direct** | SW-1, SW-2, SW-3, SW-6, SW-7, BS-2 | Scope must stay on owned pipelines/data products, not company platform ownership |
| R4 | Deliver AI/R&D PoCs, MVPs and pilots | **Direct-to-Adjacent** | BS-4 anomaly-detection PoC; FC-2 ML/NLP contribution; GN-2 PoCs; academic prototype evidence | No claim of end-to-end ownership of shared research programmes |
| R5 | Containerize, deploy and automate AI workloads | **Direct** | BS-1 Docker/Kubernetes/Ansible ML-inference integration; BS-6 automation | Do not claim model training or full ML lifecycle |
| R6 | Technical documentation and team knowledge transfer | **Direct** | BS-3 documentation/training; GN-1 training and technical ownership | - |
## Evidence Fit Breakdown
| Dimension | Weight | Score | Reasoning |
|---|---:|---:|---|
| Required qualifications | 35 | **29** | Degree, years, Python, Linux/container stack and languages are Direct; modern ML/AI-framework depth is Adjacent |
| Core responsibilities | 25 | **17** | Data pipelines, AI-workload deployment and documentation are Direct; the two leading AI/LLM-platform responsibilities are only Adjacent |
| Level and ownership | 15 | **11** | Staff + Component Owner/Application Owner exceeds generic Berufserfahrene level; role may be lateral/down-level despite meaningful build scope |
| Recency and depth | 10 | **8** | Data/Kubernetes/Python current; direct production-ML integration and deepest Linux automation date to Bosch, while current LLM evidence is configuration-level |
| Domain/tool transfer | 10 | **6** | Strong transfer from Kubernetes, Linux, data platforms and inference deployment; compute cluster, GPU operations, RAG and search are not substitutable by wording |
| Practical constraints | 5 | **3** | Thun, German, 80-100% and work authorization are excellent; compensation and project-specific PSP access remain unresolved |
| **Total** | **100** | **74** | Numerical Adjacent; strategic **Stretch** because both title-leading AI/LLM responsibilities are not Direct |
## Competitive Read
- **Obvious-fit candidate:** an MLOps/LLM-platform engineer who has built a private RAG platform, operates GPU/compute clusters, implements retrieval/connectors in Python, and can pass RUAG's PSP; ideally also Swiss military officer/NCO background.
- **Dennis's advantage:** unusually strong combination of production data engineering, Kubernetes delivery, professional Linux/Ansible automation, containerized ML inference in 24/7 manufacturing, current LLM configuration/API exposure, native German, exact Thun location and verified officer career.
- **Their advantage:** direct ownership of the two things the title names: an AI/LLM platform and production retrieval/LLM solutions, plus likely GPU cluster depth.
- **Level/scope:** Dennis is currently Staff/Engineer IV. This vacancy is labelled only Berufserfahrene, not Senior/Staff/Principal. The platform build could be substantive, but compensation and decision authority must be checked before treating it as lateral rather than a down-level move.
## Framing Strategy if the user overrides the HOLD
- **Professional identity:** Staff Data & Platform Engineer with production ML-inference integration, Kubernetes/Linux automation and a six-year officer career.
- **Strongest proof points:** BS-1, BS-6, SW-3, SW-2, SW-1/SW-7, BW-1.
- **Honest adjacent bridges:** SW-8 and LiteLLM/API exposure for LLM context; FC-2/BS-4 for AI/R&D PoCs.
- **Explicit gaps:** no owned AI/LLM platform, no RAG/retrieval implementation, no compute/GPU cluster operation, no formal LLM evaluation, no model training ownership.
- **User directives:** officer career should be visible because the JD explicitly lists it as advantageous; no invented clearance, command scope or military ICT work.
## Cover Letter Decision
- **Provisional decision: YES if proceeding.** The portal makes it optional, but the officer-career motivation and the honest transition from data/MLOps infrastructure into an AI/LLM-platform remit add information the resume cannot fully carry.
- **Verified hooks:** only the live RUAG posting: C5I support to the Swiss Army/Kommando Cyber, DevSecOps model, sovereign/high-security systems and the Thun C5I campus.
- **Do not write yet:** Phase 0 is on HOLD.
## Questions that clear the HOLD
1. **Technical scope to Marco Heinzen:** Is direct production experience implementing RAG/LLM platforms expected on day one, or is the team explicitly hiring a strong Python/Kubernetes/data-platform engineer to build that capability?
2. **Compensation to Frank Haugwitz:** What is the target base-salary range for this vacancy, including how the 13th salary and variable allowances are quoted?
3. **PSP/project access to Frank Haugwitz:** Is a German citizen with a Swiss B permit eligible for the required PSP and all intended C5I project assignments?
4. **Level:** What decision authority and technical ownership distinguish this role from RUAG's Senior/Principal engineering roles?
## Phase 2: Audience and Content Plan
### Positioning and format
- **Profile:** Swiss/DACH, German, two pages. RUAG is a traditional Swiss defence employer and the portal requests a complete dossier.
- **Professional identity:** Staff Data & Platform Engineer with production ML-inference integration, Kubernetes/Linux automation, current hands-on LLM application configuration and a six-year Bundeswehr officer career.
- **Accuracy boundary:** Do not present Dennis as an established AI/LLM-platform owner. Do not claim RAG, retrieval implementation, GPU/compute-cluster operation, model training, formal LLM evaluation or a security clearance.
- **Headline direction:** `Staff Data & Platform Engineer | Python · Kubernetes · MLOps`
- **Summary:** YES, two short sentences. Lead with current Staff-level data/platform ownership and production ML-inference deployment; add the officer career as defence-context differentiation. Describe LLM work as configuration of domain-grounded assistants, not production platform engineering.
### Skills plan, five compact lines
1. **Programming and interfaces:** Python, SQL, REST APIs; LiteLLM only as current hands-on API exposure.
2. **AI/ML engineering:** ML-inference deployment, containerized AI workloads, domain-grounded LLM assistant configuration, applied NLP/speech-recognition contribution.
3. **Platform and automation:** Linux, Docker, Kubernetes, Ansible, GitLab CI/CD, Jenkins.
4. **Data engineering:** Kafka, Airflow, PySpark, AWS S3/Glue/Athena with Iceberg/Redshift.
5. **IaC, operations and security:** CloudFormation, ELK, Grafana, Prometheus; DevSecOps tied to the 2025/2026 Security Champion assignment, not presented as a certification.
TensorFlow/Keras remain certification-context only. C++, JavaScript, Terraform, RAG frameworks and named LLM orchestration libraries are omitted.
### Planned resume bullets
| Order | Role / claim | Evidence type | Scoped bullet direction | Why it belongs |
|---:|---|---|---|---|
| 1 | Swisscom, **SW-8** | Hands-on, current | Configured domain-grounded LLM assistants in a Swisscom-owned interface by selecting available models and supplying curated knowledge for Q&A, migration assistance and data mapping. | Closest current LLM evidence; directly relevant without implying RAG or deployment ownership. |
| 2 | Swisscom, **SW-3** | Production, current; primary developer/operator | Develops, deploys and operates Python data applications on Kubernetes; automates build, test and deployment through GitLab CI/CD. | Direct match for Python, Kubernetes, containerization and automation. |
| 3 | Swisscom, **SW-7** | Production, current; scoped data-product delivery | Models and implements governed data products and source onboarding within Swisscom's company-wide AWS Data Mesh; uses Compass for metadata and lineage as a practitioner. | Shows reliable data foundations for internal AI solutions while preserving shared-platform scope. |
| 4 | Swisscom, **SW-2** | Production, current; Component Owner | Holds component responsibility for Fulfillment pipelines from Oracle/Kafka into Teradata, including governance, support, on-call duty and SLA accountability. | Strongest operations and ownership evidence. |
| 5 | Swisscom, **SW-5** | Current bounded team assignment | Serves as team Security Champion for 2025/2026 after 100 hours of DevSecOps/security training and assessment. | C5I and the posting's DevSecOps context justify the otherwise optional signal. |
| 6 | Bosch, **BS-1** | Production, historical; primary integration owner | Integrated containerized ML inference with Docker, Kubernetes and Ansible into 24/7 semiconductor production, enabling fully automated image classification. | Flagship direct evidence for deployment and automation of AI workloads. |
| 7 | Bosch, **BS-6** | Production, historical; primary administrator/automation developer | Administered and automated Linux hosts for analytics and ML workloads; developed Ansible extensions and CI-triggered configuration workflows. | Closest evidence for the platform/compute side of the target role; no cluster scale is claimed. |
| 8 | Bosch, **BS-3** | Production, historical; Application Owner | Owned lifecycle responsibilities for analytics applications and upstream pipelines, including SLOs, documentation, training and vendor coordination. | Direct operational ownership plus the JD's documentation and knowledge-transfer requirements. |
| 9 | Bosch, **BS-4** | Proof of concept | Implemented an ELK/Kafka anomaly-detection PoC with Grafana, Prometheus and Loki monitoring. | Direct evidence for AI/R&D PoCs and observability; PoC status stays explicit. |
| 10 | Fraunhofer, **FC-2** | Research-project contribution | Implemented ML/NLP components within the ARTUS research team for automatic transcription of sea-rescue communications. | Adds applied NLP and research-pilot evidence; team scope avoids sole-ownership implication. |
| 11 | Vizrt, **VZ-1 + VZ-2** | Production, historical; primary developer | Developed Python backend components for distributed video transcoding and integrated automated A/V tests as CI/CD quality gates. | Concise software-engineering and DevOps depth; avoids irrelevant C++/JavaScript skill padding. |
| 12 | Bundeswehr, **BW-1** | Verified completed service | Completed officer candidate training and officer school during six years of Bundeswehr service; left service as Second Lieutenant. | Explicit preferred qualification and the package's strongest defence-context differentiator. |
Generali remains a compact chronology entry without bullets. **Capgemini is omitted entirely** by standing user preference (six months only; short-stay impression). The Bundeswehr receives a dedicated `Militärischer Werdegang` section rather than being hidden under additional information.
### Certifications and education
- M.Eng. Computer Aided Engineering and B.Eng. Information and Telecommunication Technologies remain visible.
- Highlight AWS Certified Solutions Architect - Associate, Data Engineering with AWS and IBM AI Engineering. TensorFlow/Keras appear only within the IBM certification context.
- iSAQB CPSA Foundation may be retained if layout permits; certifications appear exactly once.
### Impact and verification
- No unverified numeric metric is planned.
- `BS-1` may use the source-backed qualitative outcome of fully automated image classification; no percentage, scale or exact workload reduction will be invented.
- Optional enrichment, not required for generation: verified adoption/users or deployment status of the `SW-8` assistants. Without it, the bullet remains configuration-level.
### Cover Letter Decision
- **YES.** The optional letter adds two things the resume cannot fully explain: the deliberate, honest bridge from data/MLOps infrastructure into an AI/LLM-platform remit, and the motivation to contribute again in a defence-related environment after six years of Bundeswehr service.
- **Narrative:** production-critical ML integration at Bosch; current Kubernetes/data/LLM-adjacent work at Swisscom; officer background and motivation to support sovereign, secure C5I capabilities in Thun.
- **Boundary:** RUAG supports the Swiss Army; employment at RUAG must not be described as rejoining the armed forces. Use `erneut im Umfeld von Landesverteidigung und gesellschaftlicher Verantwortung wirken` rather than `wieder Teil der Armee sein`.
## Phase 3/4: As-built Resume
- **File:** `output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_resume.tex`
- **Format:** German Swiss/DACH profile, A4, 11 pt, two pages, no photo.
- **As-built positioning:** `Staff Data & Platform Engineer | Python · Kubernetes · MLOps`.
- **As-built selection:** 12 bullets using `SW-8`, `SW-3`, `SW-7`, `SW-2`, `SW-5`, `BS-1`, `BS-6`, `BS-3`, `BS-4`, `FC-2`, `VZ-1 + VZ-2` and `BW-1`. Generali remains an unbulleted chronology entry. Capgemini is omitted.
- **Military evidence:** `BW-1` appears in the profile and in a dedicated `Militärischer Werdegang` section. No command responsibility, deployment, clearance or military ICT work is implied.
- **LLM boundary:** current evidence is stated as configuration of domain-grounded assistants and LiteLLM API use. The document contains no RAG, retrieval, compute-cluster, GPU-cluster, model-training or LLM-platform ownership claim.
- **Plan adjustment for layout:** certifications appear once inside `Kenntnisse`; German and English appear in the header. This preserves the information and gives page 2 a cleaner close. The Bosch title was shortened to the verified formal title to avoid a collision with the location.
- **Language sweep:** no em dash, `Baue`, `Trug`, `begrenzt`, C++, JavaScript, Terraform, RAG or fabricated orchestration framework appears. Bullets use technical, scoped verbs.
### Validation and rendering
- Canonical system preflight: **PASS, 0 warnings**.
- Document validator after final edit: **PASS, 0 warnings**.
- MiKTeX compilation: **PASS**, two runs, exactly **2 pages**.
- Log inspection: no Overfull/Underfull boxes and no LaTeX/package warnings.
- `pdftotext -layout`: employer, title, date, location and section order parse correctly across both pages.
- Visual QA: both final pages rendered to PNG at 150 dpi and inspected; no clipping, overlap, broken glyphs, isolated headings or awkward page transition.
- The compiled PDF and PNGs were temporary QA artifacts only. The deliverable remains the `.tex` source as configured.
## Post-Handover Fix and Re-Verification, 2026-08-27
The package was inherited mid-session after a usage limit on the previous model. The recorded Phase 3/4 completion was audited independently and one defect was found and corrected.
### Defect found and fixed
- **Capgemini Deutschland GmbH was present as a chronology entry.** This violates a standing user preference recorded in `CLAUDE.md` and in memory: Capgemini is omitted from all documents because the six-month tenure reads as a short stay. A corpus check confirmed this resume was the **only** generated `.tex` under `output/` containing Capgemini. The entry and its separating vertical space were removed; the visible chronology now begins at Generali, Mai 2015.
- **The Nov. 2014 to Mai 2015 gap this creates is intentional and stays.** The submitted Kdo Cy resume carries the identical gap.
- **The profile claim "über zehn Jahren" survives unchanged.** The visible record now runs Generali Mai 2015 to Aug. 2026, about eleven years three months, so the wording remains true.
### Re-verification after the fix
- Document validator: **PASS, 0 warnings** (re-run after the Thun contact-line change).
- MiKTeX compilation, two runs: **PASS**, exactly **2 pages**, **0 Overfull/Underfull boxes**, no LaTeX or package warnings. Re-run after the Thun change, this time **compiled into the output folder** rather than a temp directory.
- Extracted text of the final PDF contains **0** occurrences of "Capgemini" and reads `Thun, Schweiz` in the contact line.
- Text extraction: employer, title, date, location and section order parse correctly on both pages; footers read 1/2 and 2/2.
- Visual QA: both pages re-rendered to PNG and inspected **after** the edit. No clipping, overlap, orphaned heading or broken page transition. Page 2 closes on Ausbildung with roughly two inches of trailing whitespace, which is accepted; `CLAUDE.md` forbids page-fill quotas.
- **Note on the prior session's QA claim:** the temporary working folder still held a stale 3-page text extraction from an earlier draft, so the previously recorded visual check was made against a different document than the final one. The checks above were run against the current source.
- **Page 2 was deliberately not backfilled.** `PP-3` (private Debian server) is available as additional Linux evidence and was used in the Kdo Cy package, but re-opening the approved 12-bullet plan to fill freed space would be page-filling, not relevance.
### Contact location: RESOLVED 2026-08-27
- **The contact line was changed from `Bern, Schweiz` to `Thun, Schweiz`.** The user decided on 2026-08-27 that for roles **in Thun and its immediate surroundings** the contact line may carry Thun. This vacancy is in Thun and the user lives in Thun, so local residency is now visible to a traditional Swiss defence employer - the session's Competitive Read already counted "exact Thun location" among his advantages, and the document had been hiding it.
- **Scope of the change:** contact line only. The Swisscom position keeps `Bern, Schweiz`, because that is factually where that job is.
- **Recorded as a standing rule** in memory and `config.md`: Thun for Thun-area roles, Bern as the default everywhere else. The earlier "always Bern, ask first" note is superseded.
## User Review Round, 2026-08-27
Three changes requested by the user on review of the compiled PDF. All three applied, then re-validated, re-compiled and both pages re-inspected.
| # | Change | Basis |
|---|---|---|
| 1 | **Removed "Kein Sponsoring durch den Arbeitgeber erforderlich" from the contact line.** | User: it is common knowledge that EU citizens need no sponsorship, so the sentence adds nothing. The header still carries "Deutscher Staatsbürger" and "Schweizer Aufenthaltsbewilligung B", which resolve work authorization on their own. |
| 2 | **Bosch title changed to `Senior Data Engineer, Data Analysis (Beförderung Jan. 2021)`**, matching the Swisscom line's shape. | The promotion is **not** in `claims.json`; it is recorded LinkedIn-confirmed in `experience_bosch.md`: Data Engineer Feb 2020 - Jan 2021, **Senior** Data Engineer Jan 2021 - Dez 2022. That file also explicitly sanctions "(Senior) Data Engineer" as a display form. `claims.json` has now been corrected so this is canonical going forward. |
| 3 | **Vizrt title shortened to `DevOps Engineer`**, dropping "Test Automation /". | User request. ⚠ **Flagged accuracy note:** the canonical `official_title` for Vizrt is **"Test Automation Engineer"**; "DevOps Engineer" is half of the sanctioned `display_title` "Test Automation / DevOps Engineer". The bullet content (CI/CD quality gates, Python backend for distributed transcoding) supports the DevOps framing, and the user is the authority on his own role, but this is the one title line on the document that will not match a Zeugnis or reference check word-for-word. Recorded here so it is a conscious choice, not a drift. |
### Re-verification after the review round
- Document validator: **PASS, 0 warnings**.
- MiKTeX, two runs: **PASS**, exactly **2 pages**, **0 Overfull/Underfull boxes**, no LaTeX or package warnings.
- Both pages re-rendered and inspected. The longer Bosch title does **not** collide with `Dresden, Deutschland`; the earlier concern recorded in `CLAUDE.md` about a Bosch title/location collision does not apply at this length.
- Header now reads three lines: contact, citizenship/permit, languages.
- Build artifacts removed from the output folder; `Dennis_Thiessen_Lebenslauf.pdf` refreshed from the new build and verified identical to the `e2e_` PDF by checksum.
## Phase 5: As-built Cover Letter, 2026-08-27
- **File:** `output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_cover_letter.tex`
- **Format:** German Swiss/DACH motivation letter, A4, 11 pt, **1 page, 275 body words** (target 250-300).
- **Addressee:** Frank Haugwitz, Director Sourcing C5I-Campus, at RUAG MRO Holding AG, Business Area C5I, Uttigenstrasse 36, 3600 Thun. Submission is portal-only; the letter is uploaded, not mailed.
- **Contact location:** `Thun, Schweiz`, consistent with the resume and the new Thun-area rule.
- **Structure:** four short paragraphs plus close. (1) Role thesis: the posting pairs operating an internal AI-workload platform with building the pipelines those workloads depend on, which is the seam he already works on. (2) Bosch `BS-1` and `BS-6`. (3) Swisscom `SW-3`, `SW-2`, then the LLM boundary. (4) `BW-1` and practical fit.
### Deliberate decision: the LLM gap is named explicitly
The letter states plainly *"Eine AI-Plattform habe ich nicht aufgebaut"*, then pivots immediately to the operations/platform side and to working toward the implementation side from there.
This is a conscious departure from the Kdo Cy letter's approach, where the second-Amtssprache point was deliberately **not** raised so as not to hand the reader an objection (the Aker BP CL-A finding). The reasoning for treating this case differently:
- The AI/LLM-platform build is **R1, the title-defining responsibility and the JD's first bullet**, not a peripheral requirement. The reader will assess it whether or not the letter mentions it, so silence does not remove the objection - it only removes his framing of it.
- The session's own Cover Letter Decision names this bridge as the letter's reason to exist: the resume cannot carry it.
- The JD explicitly invites applicants who do not meet every requirement, which makes candour read as judgement rather than as a defensive gap list.
- It is one sentence, immediately followed by what he does bring. It is not a gap list.
**Reversible.** If the user prefers the Kdo Cy posture, the sentence can be cut and the paragraph still stands on `SW-3`, `SW-2` and `SW-8`.
### Claim traceability
Every claim in the letter maps to a resume bullet and a canonical ID: `BS-1` (containerized ML inference into continuous semiconductor production), `BS-6` (Linux administration, Ansible extensions, credential-caching plugin), `SW-3` (Python data applications on Kubernetes), `SW-2` (Component Owner, Oracle/Kafka to Teradata, on-call), `SW-8` (configuring domain-grounded assistants; LiteLLM API use), `BW-1` (six years, officer training, Leutnant). No new skill, tool, metric, customer or ownership claim is introduced.
### Boundaries verified in the body text
Automated scan of the letter body (comments excluded) found **0** occurrences of RAG, Retrieval, GPU, Compute Cluster, GitOps, LangChain, Terraform, fine-tuning, model training, clearance/Sicherheitsfreigabe, or "wieder Teil der Armee". The only regex hit was the substring "rag" inside *Abfragen*. **0** em dashes.
- RUAG is described as an environment of `Landesverteidigung und gesellschaftlicher Verantwortung`, never as rejoining the armed forces.
- `SW-8` stays configuration-level; no platform ownership, orchestration or evaluation is implied.
- `BS-1` claims integration, not model training.
- `BW-1` carries no clearance, command scope or military technical work.
- No unverified metric appears.
### Hook verification, 2026-08-27
All hooks are first-party, re-fetched live from the posting on the day of writing rather than trusted from the earlier retrieval.
| Claim in the letter | Evidence | Source |
|---|---|---|
| Role title `AI Engineer C5I (w/m/d)` | Verbatim match | jobs.ruag.ch posting, live 2026-08-27 |
| Legal entity `RUAG MRO Holding AG, Business Area C5I` | Verbatim match | same |
| Address `Uttigenstrasse 36, 3600 Thun` | Verbatim match | same |
| Addressee `Frank Haugwitz`, Director Sourcing C5I-Campus | Verbatim match, contact still listed | same |
| The role pairs an internal AI/LLM platform with pipelines/interfaces | `Aufbau, Weiterentwicklung und Betrieb einer internen AI-/LLM-Plattform inkl. Compute Cluster` | same |
| Posting status | LIVE, application form reachable | same |
No company-news, product or executive-statement hook was used. `cl_reference.md` prefers omitting an unnecessary hook over spending a paragraph proving company familiarity.
### Validation and rendering
- Document validator: **PASS, 0 warnings**.
- MiKTeX, two runs: **PASS**, exactly **1 page**, **0 Overfull/Underfull boxes**, no LaTeX or package warnings.
- Page rendered to PNG and inspected: single well-filled page, no clipping or awkward break.
- Build artifacts removed; `Dennis_Thiessen_Motivationsschreiben.pdf` created from the same build and verified identical by checksum.
## Critique, 2026-08-27
- **File:** `output/RUAG_AI_Engineer_C5I/critique_ruag_ai_engineer_c5i.md`
- **Evidence Fit: 74/100** - re-verified against `claims.json`, unchanged. Strategic **Stretch**; title-capability gate **FAIL** (R1 and R2 both Adjacent). PSP project access and compensation remain **UNRESOLVED**, which is independently a NO-GO trigger under `critique_framework.md` section 1. All of this was in front of the user at the 2026-08-27 override and is recorded, not re-litigated.
- **Document Quality: 85/100.** Truth and provenance **23/25** - well clear of the 8/10 automatic-failure floor.
- **Channel: Weak.** Cold portal submit; `jobs.ruag.ch` refuses email and post. Marco Heinzen's direct mobile is published and unused.
- **Scoring standard:** `critique_framework.md` governs over `SKILL.md` section 9.3/9.4 - dual score, verdicts instead of invented probabilities, no single overall score, no publications weight. `SKILL.md`'s additive coverage (domain lens, five-reader read, bridge points, 6A-6F) was kept in full.
### Verified independently, not inherited from this session file
Both documents re-validated (**PASS, 0 warnings**), recompiled twice in a scratch folder and found **byte-identical** to the shipped PDFs (133 673 B / 95 497 B), 2 pages / 1 page exactly, **0 boxes, 0 LaTeX warnings**, correct `pdftotext` order, both resume pages and the letter re-rendered at 130 dpi and inspected. Submission-named copies match their `e2e_` builds by **MD5**. `Capgemini` 0 hits; all global-forbidden patterns 0 hits.
### Full skills-token audit - the check that had not been run
Every token in the header, `Kenntnisse` block and letter was enumerated against `claims.json` `skills[].output`. **Zero `forbidden` skills on either document.** The two non-plain-`allowed` tokens present both carry their required context: **LiteLLM** (`allowed-with-context`) as "in aktueller API-Nutzung", **TensorFlow/Keras** (`certification-context-only`) only inside the Zertifizierungen line. Tokens absent from `skills[]` (Grafana, Prometheus, Loki, Jenkins, GitLab, ELK, Teradata, Oracle, Glue, Athena, Iceberg, Redshift, S3, MLOps) are all claim-backed or experience-file-backed - `skills[]` is a control list, not an inventory. **No Tier 1 truth finding anywhere in the package.**
### Tier 1 fixes - 3, none reopens the bullet plan's truth boundaries
(Tier 1 ranks by **submission impact**, not strictly by point value: T1-1 scores under cover-letter tailoring, which section 3 does not weight.)
1. **The letter drops DevSecOps, C5I, Kommando Cyber and the Army.** The JD's `Ueber den Bereich` block leads with all four and states "konsequent nach dem DevSecOps-Modell". `SW-5` is on the resume - Security Champion 2025/2026 after 100 h of Cloud Security / DevSecOps / Security by Design training - and the letter never uses it. A first-party hook matched by current canonical evidence, left on the table. One sentence closes it. **Highest-value edit in the package.**
2. **`PP-3` is missing and it is the only *current* evidence for the operating half of the role.** The letter's load-bearing sentence is *"Was ich mitbringe, ist die Betriebs- und Plattformseite dieser Rolle"*, and its resume proof is `BS-6`, which **ended Dec 2022**. `claims.json` calls PP-3 "the primary evidence for hands-on Linux administration, system hardening and security configuration, and it is current and continuous". For a defence employer screening every entrant, a decade of self-hosting *and hardening* is materially different, and it evidences Q6 better than any employer bullet. **This is a relevance argument, not page-filling** - section 7a's whitespace ban is not the basis. Kdo Cy used PP-3; this package dropped it.
3. **`governte Datenprodukte` is not a German word** - the only outright language error in the package. German IT does form participles from English verbs (*gemanagt*, *gehostet*), but *governt* is not among them, and the reader is a native-German Swiss engineer. A corpus check confirms it is a **one-off**: it appears in no other generated document, and **the user's own submitted Kdo Cy CV already carries the fix** - "Modelliere **governance-konforme** Data Products". A one-word substitution, independent of the other two, **worth taking regardless of what is decided about them**. It carried a full point of the Mechanics deduction on its own, separate from cadence.
### Tier 2 (6)
- **The headline drops "AI" from a req titled AI Engineer**, while his canonical title "Staff Data, Analytics & **AI** Engineer" contains it. Deliberate session choice, not reversed; middle option `Staff Data, Analytics & AI Engineer | Python - Kubernetes - MLOps` uses his real title and promises no platform ownership. **User's call.**
- **Four honestly closeable JD terms absent:** Data Preprocessing/Datenaufbereitung, Wissenstransfer, AI-Workloads, Containerisierung. Leave MVP/Pilotloesung - PoC is what the evidence supports.
- **`IBM AI Engineering` has no primary-source record.** The other three certs are in `thiessen_certifications.md` with issuer, date, expiry and cert number; this one is only in `bundle_ml_ai_engineer.md`. Same class as the Ansible/Linux/Bosch-promotion gaps - **verify and record, do not remove**. It matters because TensorFlow/Keras is the *sole* answer to the JD's required "moderne ML-/AI-Frameworks".
- **CL paragraph 2 is ~90 of 275 words re-telling BS-1 and BS-6**, redeemed by one genuinely new sentence. Compressing it frees the words the two Tier 1 fixes need.
- **The pivot after the gap admission concedes without evidence** - *"die Implementierungsseite wuerde ich mir von dort aus erarbeiten"* promises to learn. The gap sentence itself is a sound call and should stay; anchor the pivot in BS-1/FC-2.
- **Volunteering foreign citizenship in the letter at a PSP employer.** Already in the resume header. Lean keep (the sentence lands on "ich wohne in Thun, also am Standort selbst"), but a conscious choice.
### Tier 3
Cadence: 6/12 bullets end in a 3+ list, exactly at the section 7a ceiling, with consecutive triples at B1-B2 and B8-B9; **bullets 9 and 10 both open "Implementierte"**, a direct section 7a hit that the validator missed because its check does not cross position boundaries. `\today` in the letter - recompile if submitted after 2026-08-27. Bottom whitespace is **not** a finding (section 7a bans adding content for it). Thesis grade 1,0 is permitted; keep.
### Reader verdicts
ATS **PASS** - clean extraction. Recruiter (Haugwitz) **FORWARD** - every box he owns is above the fold. HR **COMPLETE**. Hiring manager (Heinzen) **MAYBE, and the letter decides it** - if the brief is "build our RAG stack" this is a no; if it is "run the platform while we grow the AI layer" it is a strong yes. Technical reviewer **CREDIBLE** - no claim collapses; first probes are the SW-8 boundary, compute-cluster scale, the 2008-2014 / 2009-2013 officer-track overlap, and Pull-GitOps.
### Cover letter decision
**KEEP.** The resume structurally cannot answer what this candidate thinks he is applying for given what he has not done; the letter does, in his own words, before Heinzen has to infer it. It needs T1-1 and T2-6.
## Edit 1 Baseline, 2026-08-27
- **Pages:** 2 (exact)
- **Validator:** PASS, 0 warnings
- **Overfull/Underfull boxes:** 0 | **LaTeX warnings:** 0
- **Variable bullets:** 12 (Swisscom 5, Bosch 4, Fraunhofer 1, Vizrt 1, Bundeswehr 1)
- **Skills lines:** 6 (incl. Zertifizierungen)
- **Char violations:** none - `CLAUDE.md` and `resume_reference.md` sections 4/7 replaced fixed 1L/2L/3L bands with natural-length bullets; character counts are diagnostic only and there is no page-fill quota, so the skill's "all bullets must be 2L" and "<=3 lines white space" gates do **not** apply to this project.
- **Orphan violations:** none observed in visual QA
- **White space:** ~2 in at the foot of both pages, accepted per `resume_reference.md` section 7a ("Do not add content merely to reduce bottom whitespace")
- **Cover letter:** 1 page, 275 body words, 4 paragraphs + close, validator PASS
### Edit 1 (2026-08-27): critique Tier 1 + Tier 2, user-approved
**Source:** `critique_ruag_ai_engineer_c5i.md`, user instruction 2026-08-27 ("apply 1-5, 6: use my official title, 7: cut").
**Resume**
| # | Change | Class | Source |
|---|---|---|---|
| 1 | `governte` -> `governance-konforme Datenprodukte` | MODIFY | T1-3 |
| 2 | New `Projekte` section: `Privater Debian-Server / seit rund zehn Jahren / Eigenbetrieb`, one bullet (`PP-3`) | ADD | T1-2 |
| 3 | Headline and profile opener -> **`Staff Data, Analytics & AI Engineer`**, his canonical title | MODIFY | T2-2, **user chose the official title** |
| 4 | `Wissenstransfer`, `Datenaufbereitung`, `AI-/ML-Workloads` + `Containerisierung` added where the work is real | MODIFY | T2-3 |
| 5 | Bullet 9 opener `Implementierte` -> `Baute` | MODIFY | Tier 3 (7a) |
| 6 | BS-3 list reordered to natural German: `koordinierte Hersteller, technische Dokumentation und Wissenstransfer an die Nutzer` | MODIFY | found during visual QA |
**Cover letter**
| # | Change | Class | Source |
|---|---|---|---|
| 7 | New standalone paragraph naming **DevSecOps and C5I**, carried by `SW-5` (Security Champion 2025/2026, 100 h) | ADD | T1-1 |
| 8 | Paragraph 2 compressed: the Ansible credential-plugin detail re-told resume bullet 7 | MODIFY | T2-5 |
| 9 | Pivot anchored in evidence: `erarbeite ich mir von dort aus, wie zuvor bei der ML-Inferenz in Dresden` | MODIFY | T2-6 |
| 10 | **Volunteered citizenship sentence cut** | REMOVE | T2-7, **user chose cut**. The resume header still carries `Deutscher Staatsburger` and `Schweizer Aufenthaltsbewilligung B` |
| 11 | Paragraph 3 split so the DevSecOps beat stands alone | MODIFY | found during visual QA - the merged paragraph ran to nine lines and carried three jobs |
### Two defects found during this edit that the critique had not caught
1. **The resume never loaded `babel`, so a German document was being hyphenated with English patterns.** Visible in the compiled PDF as `Hal-bleiterfertigung`, `Tran-skription` and `automa-tisierte`. Present at baseline and inherited by every earlier build. Adding `\usepackage[ngerman]{babel}` fixed the hyphenation (`Halb-leiterfertigung`, `automati-sierte`) **and** cleared the one overfull box that the longer headline had introduced. A `\hyphenation{Trans-krip-ti-on}` exception was added because ngerman still broke `Transkription` after `Tran`. **Worth checking on every future German-language package.**
2. **A one-box regression appeared and was caught by the gate, not by eye:** the longer official title reflowed the profile paragraph and overflowed by 33.6 pt. Fixed by the babel change above, not by rewording.
### Verification after Edit 1
| Gate | Resume | Cover letter |
|---|---|---|
| Canonical validator | **PASS, 0 warnings** | **PASS, 0 warnings** |
| MiKTeX, two passes | **PASS** | **PASS** |
| Pages | **2** (unchanged) | **1** (unchanged) |
| Overfull/Underfull boxes | **0** | **0** |
| LaTeX warnings | **0** | **0** |
| Submission copy refreshed + MD5 match | **yes** | **yes** |
| Visual QA at 130 dpi | both pages re-rendered and inspected | re-rendered and inspected |
- **CL body words: 295** (target 250-300). Five short paragraphs plus close.
- **Cadence improved:** list-terminated bullets **6/12 (50%) -> 4/13 (31%)**, comfortably under the 7a ceiling; **consecutive identical openers 1 -> 0**.
- Forbidden sweep across both `.tex` files and both extracted PDFs: **0 hits** for Capgemini, LangChain/LangGraph/LlamaIndex, Terraform, Azure, GCP, GitOps, TypeScript, FastAPI, Flask, Django, every global-forbidden pattern, and the standalone acronym `RAG` (the only `rag` matches are inside *Fragen* and *Abfragen*). **0 em dashes.**
### Baseline comparison
| Metric | Before | After | Delta |
|---|---:|---:|---|
| Pages | 2 | 2 | 0 |
| Bullets | 12 | 13 | +1 (`PP-3`, inside the 11-14 guide) |
| Boxes | 0 | 0 | 0 |
| Validator warnings | 0 | 0 | 0 |
| List-terminated bullets | 50% | 31% | **-19 pts** |
| Consecutive identical openers | 1 | 0 | **-1** |
| CL body words | 275 | 295 | +20 (in range) |
| German hyphenation | English patterns | **ngerman** | fixed |
### KB corrections made during this edit
- **`IBM AI Engineering` verified and recorded.** The user supplied the certificate; it was read directly. IBM via Coursera, **4 Jun 2020**, no expiry, verify code `3ZBZFVAL6A34`, name on certificate `Dennis Thiessen` (ss/sz variant, as on ITIL and iSAQB). Added as entry **#8** in `thiessen_certifications.md`. **T2-4 is closed** - the resume's `IBM AI Engineering mit TensorFlow/Keras` line is now backed by a primary source, which matters because it is the sole evidence for the JD's required *moderne ML-/AI-Frameworks*.
- **`PyTorch` upgraded in `claims.json`** from `coursework-or-personal-unverified` to `evidence: certification`. The same certificate includes *Deep Neural Networks with PyTorch*. `output` stays `certification-context-only`, so this changes nothing on any current document - it just means PyTorch now has a real source if a JD ever asks. Patched as text, not re-serialized, per the CRLF/compact-JSON note in `CLAUDE.md`.
- The certificate is explicitly **non-credit** and confers no grade or degree; never present it as an academic qualification.
## Critique Round 2, 2026-08-27 (post-Edit 1)
- **File:** `critique_ruag_ai_engineer_c5i.md`, rewritten as Round 2 with a Round 1 resolution table.
- **Evidence Fit: 74/100 - unchanged.** Strategic Stretch, title-capability gate **FAIL**, PSP and compensation **UNRESOLVED**. Document edits cannot move candidate-role fit. **`PP-3` deliberately did not move it**: it is a personal project and cannot convert an Adjacent title-capability into a Direct one - counting it would be the "bridge valued as highly as direct evidence" error `critique_framework.md` section 5 warns against.
- **Document Quality: 85 -> 93/100.** Truth and provenance **23 -> 24/25**.
- **Channel: Weak - unchanged.** Heinzen's published direct line still unused.
- **Tier 1 findings: NONE.** No truth, fit or relevance defect remains that editing can fix.
### Dimension movement
| Dimension | R1 | R2 | Why |
|---|---:|---:|---|
| Truth and provenance | 23 | **24** | IBM cert primary-source verified; new PP-3 and SW-5 content audited clean; headline now his real title, not a positioning label |
| Information hierarchy | 17 | **19** | Headline carries the req's word above the fold; `Projekte` parses cleanly |
| Bullet evidence and impact | 16 | **16** | **Unchanged - now the weakest dimension** |
| Relevance and terminology | 12 | **15** | 22/22 claimable JD terms; all six gap terms still correctly absent |
| Skills evidence | 9 | **9** | Unchanged; `MLOps` remains an umbrella term with no `skills[]` entry |
| Mechanics/readability | 8 | **10** | `governte` fixed, German hyphenation fixed, cadence 50%->31%, consecutive openers 1->0 |
### Reader verdicts
ATS **PASS**. Recruiter **FORWARD** - the Round 1 friction point is gone, the headline now shows the req's own word. HR **COMPLETE**. Hiring manager **still MAYBE, but a stronger MAYBE** - the underlying R1/R2 split is unchanged; the letter now argues the operating-layer case more completely. Technical **CREDIBLE**, with a new and welcome probe invited by PP-3: "what have you actually hardened, and how do you patch it?"
### The one dimension that did not move
**Bullet impact: 12 of 13 bullets describe activity with no outcome clause; only `BS-1` says what changed.** Every other claim in `claims.json` carries `metrics: unverified`, and inventing figures is barred - so the resume is correctly written but flat where a hiring manager looks for consequence. **The fix is upstream, not in this package:** if any verified outcome exists for `SW-2` (incident/SLA record), `SW-3` (deployment frequency, service count), `SW-7` (data products delivered, sources onboarded) or `BS-3` (applications owned), recording it in `claims.json` would lift this dimension for **every** future package. Do not invent anything to close it here.
### Remaining Tier 2 / Tier 3
Officer career still on page 2 (structural; moving it breaks reverse chronology). Headline now duplicates the Swisscom title line verbatim (cosmetic). The CL's *"entspricht damit meiner heutigen Praxis"* is a soft alignment claim worth being ready to answer, not worth rewording. Bosch bullet lengths cluster at a 5-word spread (validator passes; diagnostic only). `\today` - recompile if submitted after 2026-08-27.
## SUBMITTED 2026-08-27
Application sent via `jobs.ruag.ch` (portal only - email and postal applications are explicitly refused).
Application ID **18027**, publish ID 10075372.
### As-submitted record
| Field | Value |
|---|---|
| Employer | RUAG MRO Holding AG, Business Area C5I |
| Role | AI Engineer C5I (w/m/d), Thun, 80-100% |
| Submitted | **2026-08-27** |
| Fit class | **Stretch** (numerically top-of-Adjacent at 74) |
| Evidence Fit | **74/100** |
| Hard gate | **FAIL** - R1 and R2, the two title-defining responsibilities, are Adjacent not Direct |
| Document Quality | **93/100** (85 pre-edit), truth and provenance 24/25, **0 Tier 1 findings** |
| Channel | **Weak** - cold portal submit |
| Cohort | evidence-first-2026-01, now **6/10**; **the sole Stretch slot is CONSUMED (stretch 1/1)** |
| Outcome | applied - awaiting response |
**Submitted documents:** `Dennis_Thiessen_Lebenslauf.pdf` (2 pages, 13 bullets) and
`Dennis_Thiessen_Motivationsschreiben.pdf` (1 page, 295 words, 5 short paragraphs). Both validators PASS with
0 warnings, 0 boxes, clean two-pass compiles, correct text-order parse and visual QA of every page. Submission
copies were MD5-verified identical to their `e2e_` builds before sending.
### What went in, in one line
Staff Data, Analytics & AI Engineer - his canonical title - with production ML-inference integration (`BS-1`),
professional Linux/Ansible automation (`BS-6`), current Kubernetes and data-product work (`SW-3`, `SW-7`, `SW-2`),
current self-hosted Linux and hardening (`PP-3`), DevSecOps practice tied to C5I's stated operating model (`SW-5`),
and the canonical six-year Bundeswehr officer career (`BW-1`). The letter names the AI-platform gap in one sentence
and pivots to the operations and platform side.
### Live screening topics - none of these were resolved before sending
1. **Compensation.** No band published. Kununu samples are small and non-authoritative but suggest a material cut
against the ~CHF 140k current and the 180k bar. The Thun-local exception permits near-current pay, not a large
reduction. **Ask Frank Haugwitz.**
2. **PSP and project access.** RUAG screens every entrant. No citizenship requirement is stated and German
citizenship plus a Swiss B permit is not an evidenced disqualifier, but role-specific project eligibility is
unknown. **Ask Frank Haugwitz.**
3. **Level.** The req is labelled only "Berufserfahrene" and may be lateral or a down-level move against current
Staff + Component Owner scope. **Ask what decision authority separates this from RUAG's Senior/Principal roles.**
4. **The question the application actually turns on:** does C5I want someone who has *built* an AI/LLM platform, or
someone who can *operate* one and grow into building it? **Marco Heinzen, +41 79 568 14 96** - the published
direct line that was never used. A call now would still upgrade the channel from Weak to Moderate.
### Honest gaps that travel with this application
No AI/LLM platform built or operated; no compute or GPU cluster; no RAG, retrieval or document-solution
implementation; no formal LLM evaluation; no model-training ownership. ML/AI-framework depth is certification-context
only (IBM AI Engineering, now primary-source verified). These are real and were never written around.
### Standing note for any future RUAG C5I application
The KB now holds a verified, live JD and a full Phase 0 for this business area, and RUAG C5I has **five other reqs**
already in the scout log (Senior DevOps Engineer C5I and Data Lakehouse / Senior Data Platform Engineer are both
shortlisted and are **materially better fits than this one** - they sit on the DevOps/data-platform lane where his
evidence is Direct rather than Adjacent). If this application is rejected, that is **not** a signal to drop RUAG;
it is a signal to target the reqs that match the evidence. Note also that the cohort's Stretch slot is now full,
so any further RUAG application must clear Core or Adjacent on its own merits.
## Output Files
- Verbatim JD: `output/RUAG_AI_Engineer_C5I/JD_ruag_ai_engineer_c5i.txt`
- Session: `output/RUAG_AI_Engineer_C5I/session_ruag_ai_engineer_c5i.md`
- Resume: `output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_resume.tex`; finished, Capgemini defect corrected, re-validated and re-compiled 2026-08-27; awaiting user approval.
- Compiled resume PDF: `output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_resume.pdf` and a submission-named copy `output/RUAG_AI_Engineer_C5I/Dennis_Thiessen_Lebenslauf.pdf`, both built 2026-08-27 from the corrected source. ⚠ **`Dennis_Thiessen_Lebenslauf.pdf` is a copy, not a build target.** pdflatex only regenerates the `e2e_` file, so the submission-named copy must be **refreshed after every recompile** or it silently becomes the old version under the exact filename that would be attached to the application.
- Cover letter: `output/RUAG_AI_Engineer_C5I/e2e_ruag_ai_engineer_c5i_cover_letter.tex`, with `e2e_ruag_ai_engineer_c5i_cover_letter.pdf` and the submission-named `Dennis_Thiessen_Motivationsschreiben.pdf`. Built 2026-08-27; awaiting user approval.
- Critique: `output/RUAG_AI_Engineer_C5I/critique_ruag_ai_engineer_c5i.md` - **CURRENT**, written 2026-08-27. Evidence Fit 74, Document Quality 85, Channel Weak.
## Status
- **Phase 0:** COMPLETE 2026-08-27.
- **Fit gate:** Explicitly overridden by the user on 2026-08-27; strict hard-gate FAIL, Evidence Fit 74 and strategic Stretch remain recorded.
- **Phase 2:** COMPLETE; audience and 12-bullet content plan prepared.
- **Phase 3/4:** COMPLETE; resume generated, validated, compiled and visually inspected. **Re-audited after handover 2026-08-27:** one standing-preference defect (Capgemini) found and fixed, then re-validated, re-compiled and re-inspected. See Post-Handover Fix and Re-Verification.
- **Resume:** finished; awaiting user approval.
- **Phase 5 (cover letter):** COMPLETE 2026-08-27; generated, hooks verified live, validated, compiled and inspected.
- **Cover letter:** DONE; awaiting user approval.
- **Application/outcome:** **SUBMITTED 2026-08-27** via jobs.ruag.ch, application ID 18027. Awaiting response. Cohort recorded: Stretch slot consumed, cohort 6/10.
- **Contact location:** RESOLVED 2026-08-27 - `Thun, Schweiz`.
- **Resume:** approved by the user 2026-08-27 (implicitly, by invoking `/make-cl`), after three review changes: sponsoring sentence removed, Bosch title with promotion date, Vizrt shortened to DevOps Engineer.
- **Critique:** **CURRENT (Round 2, 2026-08-27, post-Edit 1)** - Evidence Fit **74/100**, Document Quality **93/100**, Channel **Weak**, **Tier 1 findings: none**. Round 1 (pre-edit) scored - Evidence Fit **74/100**, Document Quality **85/100**, Channel **Weak**, **no Tier 1 truth finding**, 3 Tier 1 fixes open (DevSecOps hook in the letter, PP-3 on the resume, `governte` -> `governance-konforme`).
- **Edit 1:** COMPLETE 2026-08-27. All three Tier 1 fixes applied, plus Tier 2 items 3-5 and the two user decisions (official title; citizenship sentence cut). Both documents re-validated, recompiled, re-rendered and inspected; submission-named PDFs refreshed and MD5-matched.
- **Next:** await response. Optional and still worthwhile: call **Marco Heinzen, +41 79 568 14 96** about technical scope and level, and **Frank Haugwitz** about compensation and PSP/project eligibility - all three are now live screening topics rather than pre-cleared ones. Do not react to a single rejection with positioning changes (`application_strategy.md` sections 38/74); the better-fitting RUAG C5I DevOps and Data Platform reqs are already shortlisted in the scout log. **Unchanged and unaffected by any edit:** Evidence Fit 74, Stretch, title-capability gate FAIL, and the open PSP / compensation / level questions. The highest-value remaining action is not a document change - it is the call to Marco Heinzen (+41 79 568 14 96). Compensation, PSP/project access and role level remain open practical questions and are unaffected by the documents. Because the cover-letter decision is YES, proceed with `/make-cl` after approval. Compensation, PSP/project access and level remain open practical questions.
output/SBB_DataEngineer_AssetMgmt/JD_SBB_DataEngineer_AssetMgmt.txt
+83
View File
@@ -0,0 +1,83 @@
Data Engineer (m/w/d) - Asset Management Infrastrukturanlagen
Jetzt bewerben
Beweg die Schweiz mit uns.
Data Engineer (m/w/d) - Asset Management Infrastrukturanlagen
Bern / Work Smart, ab sofort oder nach Vereinbarung, 60-100%
Jetzt bewerben
Darauf kannst du dich freuen.
Benefits wie kostenloses GA, vergünstigtes Reisen im Ausland - auch für die Familie - findest du auf unserer Website unter Benefits. 
Das kannst du bewegen.
In dieser spannenden Rolle baust du Datenpipelines auf, betreibst und entwickelst sie weiter auf der Enterprise Analytics Plattform der SBB, die Batch- und Streaming-Ansätze vereint.
Unterschiedlichste Datenquellen integrierst du via APIs/Kafka und ermöglichst damit Analysen, Dashboards sowie vorausschauende Wartung.
Datenprodukte und semantische Modelle gestaltest du auf Snowflake und Power BI.
Du entwickelst RAMSI weiter – von der Webanwendung (Java Spring Boot/Angular) über Reports bis hin zu Dashboards für unsere Fachspezialist:innen.
In enger Zusammenarbeit mit dem Business nimmst du Anforderungen auf und sorgst für nachhaltige Lösungen im Betrieb.
Dein Wissen und deine Erfahrung teilst du im Team und trägst zur fachlichen Weiterentwicklung bei.
Das bringst du mit.
Starke Kommunikationsfähigkeiten und Teamorientierung mit der Bereitschaft, gemeinsam mit unseren Fachspezialist:innen konkrete Lösungen zu erarbeiten und die fachliche Entwicklung anderer Dateningenieur:innen zu fördern.
Agile, selbständige und lösungsorientierte Arbeitsweise mit ausgeprägter Verlässlichkeit und Bereitschaft, Verantwortung für die Qualität und Nachhaltigkeit technischer Lösungen zu übernehmen.
Mehrjährige praktische Erfahrung im Entwurf, Aufbau, der Entwicklung und dem Betrieb cloudbasierter Datenplattformen mit modernem Data Stack (ETL/ELT, Data Warehouses, Streaming, Orchestrierung – z. B. Snowflake, DBT, Helm, Argo Workflows, Kafka).
Fähigkeit, skalierbare und kosteneffiziente Datenarchitekturen zu entwerfen und Trade-offs zwischen Latenz, Konsistenz und Kosten kritisch zu beurteilen.
Erfahrung in Datenmodellierung (Star-/Snowflake-Modelle) und Reporting mit Power BI, fortgeschrittene SQL-Kenntnisse sowie Kenntnisse zur Abfrageleistungsoptimierung.
Software-Engineering-Fähigkeiten in Python, inklusive Tests, CI/CD, Packaging- und Deployment-Best-Practices.
Dein Lohn und deine Benefits.
Wir haben ein faires und fortschrittliches Lohnsystem und setzen uns aktiv für Lohngleichheit ein. Deshalb sind weiterführende und detaillierte Informationen zum Lohnsystem bei der SBB transparent einsehbar. Diese Stelle befindet sich im Anforderungsniveau K, Regionalzulage Stufe 1. Mehr zu Lohn und Benefits.
Damit gelingt der Einstieg.
Dass du mit uns die Schweiz bewegen willst, ist für uns die grösste Motivation. Darum verzichten wir bei dieser Stelle auf ein Motivationsschreiben. Falls du uns deine Motivation dennoch mitteilen möchtest, hast du in unserem Bewerbungstool nach wie vor die Möglichkeit dazu.
Grosses machen ist einfacher per du. Das beginnt bereits bei deiner Bewerbung.
So geht's weiter.
Bewerbung
Virtuelles Kennenlernen mit HR und Führungskraft
Persönliches Kennenlernen
Finaler Entscheid
Noch Fragen?
Zum Bewerbungsprozess
SBB Recruiting Support
Telefon + 41 51 220 20 29
recruiting@sbb.ch
#einfachDu
FAQ zur Bewerbung
Zur Stelle
Job ID 103755
Andri Wienandts, People Leader
Telefon +41 79 364 62 53 
Bewerbungen per E-Mail können leider nicht berücksichtigt werden.
Für diese Stelle nehmen wir keine Dossiers von Personalvermittlern entgegen.
Das bieten wir.
Benefits
Wo immer möglich setzen wir auf flexible Arbeitszeiten und moderne Arbeitszeitmodelle. Darüber hinaus sind wir als Arbeitgeberin auch flexibel, wenn es darum geht, dein Berufsleben mit deinem Privatleben in Einklang zu bringen.
Mehr erfahren
Unterschiede willkommen
Mit über 150 verschiedenen Berufen bieten wir eine beeindruckende Vielfalt an Einstiegs- und Entwicklungsmöglichkeiten. Ebenso vielfältig sind unsere Mitarbeitenden. Darum engagieren wir uns dafür dass alle Mitarbeitenden ihre individuellen Stärken und Denkweisen einbringen können.
Mehr erfahren
Karriere und Entwicklung
Wir fördern unsere Mitarbeitenden entlang ihren persönlichen Bedürfnissen und Ambitionen. Dafür stellen wir ein breites Angebot an fach- und bereichsübergreifenden Weiterbildungsangeboten zur Verfügung. An ausgewählten externen Weiterbildungen, beteiligen wir uns finanziell und zeitlich.
Mehr erfahren
Wir bewegen heute schon Grosses.
Jetzt bewerben
Grosses bewegen.
output/SBB_DataEngineer_AssetMgmt/comp_worksheet_sbb.md
+80
View File
@@ -0,0 +1,80 @@
# Total-Comp Worksheet — SBB vs Swisscom
**Purpose:** compute the break-even *before* the call on 2026-09-09, so that when they name a figure you know immediately whether it clears — instead of working it out afterwards.
**Honesty note on this file:** I could not retrieve the Anforderungsniveau K figures — they are not public, which is exactly why Q1 exists. Anything below marked ⚠️ is an estimate or a recollection you should verify yourself rather than trust. The Swisscom column is mostly blank because I only hold the ~140k all-in headline; you know the split.
---
## 1. The Swisscom baseline (fill from your Lohnausweis / total-rewards statement)
| Component | Amount CHF/yr | Notes |
|---|---:|---|
| Base salary (12×) | | |
| 13th month | | in-band or on top? |
| Bonus / variable (actual, not target) | | use a 3-year average, not the good year |
| Employer pension contribution | | **from the Lohnausweis / PK statement, not the deduction you see on the payslip** |
| Other (allowances, RSU, phone, insurance) | | |
| **Total all-in** | **~140 000** | the figure of record in `config.md` |
| Weekly hours | | |
| Vacation days | | |
| Commute cost Thun↔Bern (what you actually pay today) | | this becomes a *saving* on the SBB side |
---
## 2. The SBB side (fill live, on the call)
| Component | Amount CHF/yr | How to get it |
|---|---:|---|
| Base, Anforderungsniveau **K** | | **Q1.** Ask for the band *and* the likely entry point for your profile |
| Regionalzulage **Stufe 1** | | stated on the posting; ask for the franken amount |
| 13th month | | ask explicitly whether it is inside or on top of the band figure |
| Employer pension contribution (PK SBB) | | ask for the **employer** %, and which plan/Sparplan |
| **Free GA** | ⚠️ ~4 000 (2nd cl.) / ~6 500 (1st cl.) | **verify the current list price yourself.** Ask which class, and how it is taxed — a staff GA is declared on the Lohnausweis, so the net benefit is below list |
| Family international travel discount (FVP) | | worth something real if you travel; hard to price — estimate honestly or score it qualitatively |
| Kidz Care childcare subsidy | | **Q4.** Scales on *gross household* income; bracket table is intranet-only. Assume well below the headline 90% |
| Commute saving (from §1) | | the GA makes Thun↔Bern free — count it once, here or in the GA line, **not both** |
| **Total all-in** | | |
| Weekly hours | | |
| Vacation days | | |
| Pensum offered (60–100%) | | a 90% option changes the CHF/hour maths materially |
---
## 3. Break-even
Compute this **before** the call and write the number here:
> **Minimum acceptable SBB base = ______ CHF**
>
> derived so that: base + 13th + Regionalzulage + employer pension + GA (net of tax) + commute saving ≥ Swisscom all-in
Then on the call you are comparing one number to one number.
**Three traps that make this comparison lie:**
1. **Base-to-base is meaningless here.** The GA, the Regionalzulage and the pension employer share are all real money that never appears in a base-salary comparison. Conversely a Swisscom bonus is real money SBB will not match.
2. **Count the commute once.** Free GA and "commute saving" are the same franken. Double-counting flatters SBB by four figures.
3. **The GA is taxed.** It goes on the Lohnausweis, so the net value is below list price. Don't book the sticker number.
---
## 4. What this is really deciding
The band is capped and collectively bargained — **there is no negotiation lever at the top of K** the way there is on a private offer. So this is a yes/no against a fixed number, not an opening position.
Set your own line before the call:
- **Clears break-even** → the level question (Q2) becomes the deciding one, not the money.
- **Below break-even but within ~10%** → this is the Bern/Thun WLB-exception trade you already defined: below-bar comp accepted for a local seat with a real mission. Worth it only if Q2 comes back with genuine design scope. It has never been worth it for a *below-level* seat — that is why BKW was declined.
- **Well below** → decline cleanly, but still finish the call and ask everything. See §5.
---
## 5. Why the call is worth doing even if the number fails
- **It is the first real data point on what the entire Bern/Thun local tier pays.** RUAG, Swissgrid, BKW and BFH all sit in comparable Swiss public-sector/parapublic band structures. Whatever K turns out to be, it calibrates that whole lane — which is otherwise pure guesswork.
- **It is interview practice with nothing at stake.** NATO JWC (final interviews 2nd half of October), BIS Basel and Microsoft Principal FDE are all still live. You do not want one of those to be your first interview in years.
- **Not wanting it makes you better at it.** You can ask the blunt questions — band, scope, RAMSI share — without hedging, which is exactly how they should be asked anyway.
Record the K figure in this file afterwards regardless of outcome. It is reusable intelligence for the whole local tier.
output/SBB_DataEngineer_AssetMgmt/critique_sbb_data_engineer_asset_mgmt.md
+166
View File
@@ -0,0 +1,166 @@
# Critique — SBB, Data Engineer (m/w/d), Asset Management Infrastrukturanlagen
**Date:** 2026-08-21 · **Job ID:** 103755 · **Package:** 2-page English resume, no cover letter (SBB waives it)
**Documents reviewed:** `e2e_sbb_data_engineer_asset_mgmt_resume.tex` → `.pdf` (2 pages)
**JD:** `JD_SBB_DataEngineer_AssetMgmt.txt` — verbatim Playwright scrape 2026-08-21, live posting. **JD integrity: PASS.**
> **Framework note.** `.claude/skills/critique/SKILL.md` Part 3 asks for a single 8-dimension score including a 10% Publications weight. That conflicts with `critique_framework.md` ("Never collapse candidate-role fit and document polish into one score") and with CLAUDE.md ("Record Evidence Fit, Document Quality, and Channel Strength separately; never collapse them into one optimistic score"). **CLAUDE.md and `critique_framework.md` govern**, so this critique reports three separate scores. The Publications dimension is not applicable — this is a resume for an industry role and Dennis has no publications. The skill's genuinely useful lenses (reader sequence, tiered fixes, interview bridges, mechanical verification) are retained below.
---
## 1. Fit verdict and hard-gate audit
**Hard gate: PASS.** No minimum qualification is a Gap. The title-defining capability — build and operate cloud data pipelines and governed data products — is Direct and current. No clearance, relocation, language or authorization issue.
**Evidence Fit: 79/100 — Core (lower end).** Unchanged from Phase 0; a document cannot change candidate-role fit. Full dimension table is in the session file. The binding constraint is not fit but **level**: this seat sits below current Staff + Component Owner scope (scored 9/15), the same shape declined at BKW.
**Unresolved practical constraint:** Anforderungsniveau K is a capped GAV band with no published figures. Not a fit failure, but it is the decision gate.
---
## 2. Document Quality: 90/100
| Dimension | Weight | Score | Assessment |
|---|---:|---:|---|
| Truth and provenance | 25 | **24** | Every bullet traces to a canonical claim ID. Validator PASS (0 warnings). SW-1 scoped to "pipelines in the Fulfillment and Product Analysis domains" + "contributing to the wider company migration programme"; SW-7 builds *data products* **within** the mesh, never the mesh. Compass and Spotfire correctly scoped as practitioner and co-owner. No unverified metric anywhere. |
| Information hierarchy | 20 | **19** | Employer bold first, dates on the same line, title and location second, promotion shown. The guarded page break keeps every page's content under its own header. Minus one: the M.Eng. entry wraps onto two lines in the FIXED Education block, pushing its date to the second line. |
| Bullet evidence and impact | 20 | **16** | Bullets are specific, scoped and verb-accurate. **Zero quantified outcomes anywhere in the document** — the correct call given every canonical metric is `unverified`, but it is still a real cost against competitors who will quantify. Bullet 5 carries two accomplishments (delivery; automation + RCA) — at the limit, not over it. |
| Relevance and terminology | 15 | **11** | Covers most JD vocabulary naturally. Three cheap, *honestly available* terms are missing (see Tier 1/2). The largest terms cannot be carried at all. |
| Skills evidence | 10 | **10** | Six lines, every entry canonical `allowed` or `allowed-with-context`, every one interview-ready. No self-ratings. No tool listed merely because the JD says it. |
| Mechanics and readability | 10 | **10** | Compiles to exactly 2 pages. No clipping, overlap, orphaned heading or bad break. Umlauts render correctly. Bullets run 20–31 words with natural variation; none exceeds 35. |
| **Total** | **100** | **90** | Strong. Truth/provenance is well above the 8/10 automatic-failure line. |
---
## 3. Channel Strength: **Weak** — but uniquely upgradeable
Currently a tailored cold application, same as every entry in the log. **This is the first posting in the entire history that publishes a named hiring contact with a direct mobile number** (Andri Wienandts, People Leader, +41 79 364 62 53). One call moves this to Moderate. Given that cold-channel outcomes have been uniformly poor — Google 1 day, Microsoft ISE 39 days, Equinor 79 days, all no-interview — the channel is a higher-leverage variable than any remaining document edit.
---
## 4. ATS / parser read
Extraction is clean; the layout is conventional and machine-readable. Term coverage against the JD:
**Present and natural:** ETL/ELT, Kafka, Airflow, streaming, data modelling, SQL, Python, tests, CI/CD, packaging, deployment, data products, Kubernetes, Docker, data governance, data quality, on-call, production operation, stakeholders, training, Teradata, Oracle, AWS, anomaly detection.
**Absent and honestly addable — the actionable finding:**
| Term | Status | Canonical support |
|---|---|---|
| query performance optimization | **0 occurrences** | BS-2's 3L variant: "optimized query performance and data availability". The JD names it in a *required* line. |
| agile | **0 occurrences** | SW-3 and SW-4 both record "agile DevOps team". The JD's second required line opens with "Agile ... Arbeitsweise". |
| API | **0 occurrences** | The JD says "Datenquellen integrierst du via APIs/Kafka". Kafka is covered; APIs are not. SW-7 onboards source systems. |
**Absent and NOT addable — accepted risk:** Snowflake, dbt, Argo Workflows, Helm, Power BI, Spring Boot, Angular, *predictive maintenance*, *asset management*. An automated screen keyed on any of these rejects the document, and nothing honest changes that. Note "asset management" appears once in a source-file **comment** only, not in rendered text.
> **On "predictive maintenance" specifically:** the phrase is deliberately absent. The nearest honest evidence is anomaly detection on streamed logs (PoC), defect classification, and a 2013 master's thesis on vibration-based condition monitoring that `claims.json` explicitly forbids presenting as a validated system. The Education block does carry the thesis title — *Development of a Web-Based Remote Fault Diagnosis System* — which is factual and does real work here without any claim being made. **Do not insert the term.**
---
## 5. Reader sequence
| Reader | Verdict | Reason |
|---|---|---|
| **Parser / ATS** | Pass, with one exposure | Clean extraction and strong general term coverage; fails any screen keyed on the three named products. |
| **Recruiter (10s)** | **Forward** | Bern-local Staff engineer at Swisscom, German native, exact title match, no permit question. Nothing to disqualify. |
| **Hiring manager (2min)** | **Interview — probable** | The lead bullet is Component Ownership with on-call, quality and governance. That is this JD's "Betrieb" line answered by someone doing it today at a peer Swiss enterprise. |
| **Technical reviewer (10min)** | **Credible** | Expect two challenges: (1) "You have not used Snowflake or dbt — map your Teradata/Glue/Athena work onto them." (2) "How would you model a semantic layer over asset condition data?" Both are answerable; neither is bluffable. |
**Ceiling analysis.** The document is not the limiting factor — Document Quality 90 with a clean provenance record. The ceiling is set by (a) the three unnameable products, (b) the level inversion, and (c) a cold channel. Only (c) is under your control this week.
---
## 6. Competitive read
The obvious-fit competitor is a Swiss data engineer with 5+ years on the exact named stack, ideally already inside a transport, utility or infrastructure operator, and willing to maintain a Java/Angular internal web application. That last trait thins the field considerably — most pure data engineers do not want it.
**Dennis's real advantage:** he performs the operational half of this role today, in the same city, at comparable enterprise scale — plus a combination that is genuinely rare for this particular req: industrial sensor and asset data (defect records, wafer inspection images, PCM electrical parameters), anomaly detection over streamed logs, and containerized delivery into a continuously operating 24/7 environment.
**Their real advantage:** named-product fluency, and rail-domain vocabulary (asset taxonomy, EN 50126 RAMS, the swissTAMP/RIS ecosystem).
---
## 7. Claim audit
| Claim | Canonical ID | Direct/Adjacent | Safe? | Note |
|---|---|---|---|---|
| "Own Fulfillment ETL pipeline components as Component Owner" | SW-2 | Direct | ✅ | Allowed verb "own"; object scoped to components. |
| "Migrated pipelines in the Fulfillment and Product Analysis domains… contributing to the wider company migration programme" | SW-1 | Direct | ✅ | Scoped object + hedged programme reference. Avoids all four SW-1 forbidden phrasings. |
| "Build and model governed data products within Swisscom's company-wide Data Mesh" | SW-7 | Direct | ✅ | Verbs "build"/"model"/"onboard"/"maintain" all allowed; mesh is the container, never the object. |
| "active metadata and lineage (Atlassian Compass)" | SW-7 | Direct | ✅ | Practitioner framing; no admin/rollout/ownership claim. |
| "Build and operate Python data applications on Kubernetes with GitLab CI/CD" | SW-3 | Direct | ✅ | |
| "Integrated containerized ML inference… replacing manual image-based defect classification" | BS-1 | Direct | ✅ | "Integrated" not "trained"; no model-development claim. |
| "Built an anomaly-detection proof of concept" | BS-4 | Adjacent | ✅ | **PoC label retained** as required. |
| "Co-owned the TIBCO Spotfire environment… co-presented" | BS-5 | Direct | ✅ | Co-ownership and co-presentation both preserved. |
| "Served as Application Owner… defining SLOs" | BS-3 | Direct | ✅ | |
| Summary: "Component Ownership covering data quality, governance and on-call operation" | SW-2 | Direct | ✅ | No scale or impact inflation. |
**No Tier 1 truth findings.** Security Champion correctly omitted. No LOC or test counts. No fabricated tools.
---
## 8. Tiered improvements
### Tier 1 (≥1 pt) — ✅ RESOLVED 2026-08-21
1. ~~**Add query-performance optimization to BS-2.**~~ The JD names "Kenntnisse zur Abfrageleistungsoptimierung" in a *required* line and the document was silent on it, despite canonical support in BS-2's 3L variant. **FIXED:** BS-2 now reads "…tuning query performance for analysis teams working with semiconductor manufacturing data…", and "query performance tuning" was added to the Pipelines skills line. Two occurrences where there were zero. Validator re-run PASS, still 2 pages.
### Tier 2 (0.3–0.9 pt)
2. ~~**Surface "agile"**~~ — ✅ **RESOLVED 2026-08-21.** SW-3's canonical 2L variant records "in an agile DevOps team"; that phrase is now in the SW-3 bullet. The JD's second required line opens with *Agile … Arbeitsweise*.
3. **Name APIs alongside Kafka** in the SW-7 onboarding bullet, matching the JD's "via APIs/Kafka". **NOT APPLIED — deliberately.** `claims.json` SW-7 records "onboard source systems" with no mention of APIs, and SW-2's ingestion is Oracle → Kafka. There is no canonical evidence that source onboarding used APIs, so adding the term would be vocabulary substitution, exactly what `ai_fingerprint_rules.md` §5 forbids. **Only add if Dennis confirms from memory that it is true.**
### Tier 3 (<0.3 pt)
4. The M.Eng. Education entry wraps to two lines, orphaning its date. FIXED section — flagged, not edited.
5. Consider whether "Staff Data Engineer" in the headline should read "Staff Data, Analytics & AI Engineer" to match the employer's formal title exactly. Current form is a permitted normalization and reads better against this JD; no change recommended.
---
## 9. Interview bridge points
1. **Snowflake / dbt.** "I have not used them by name. I have run the same shape of work on Teradata and Oracle, then on Glue, Athena with Iceberg and Redshift, with Airflow for orchestration — modelling, testing and lineage discipline transfer; the syntax is what I would pick up."
2. **Power BI.** "I co-owned a BI platform rather than just consuming one — TIBCO Spotfire at Bosch, including C# extensions and wafer-map visualizations, which I co-presented at TIBCO Analytics Forum 2022."
3. **Predictive maintenance.** Lead with Bosch sensor and process data and the ELK/Kafka anomaly-detection PoC. The master's thesis on vibration-based condition monitoring is a genuine interest signal — **describe it as a methods prototype, never as a validated system.**
4. **RAMSI / Spring Boot / Angular.** Be straightforward: Java is professional but historical, Spring Boot was a PoC contribution roughly a decade ago, Angular is absent. Better to ask what share of the role this actually is than to oversell it.
5. **The level question, asked of them.** "The posting reads as a senior IC seat. Does it carry design ownership for the pipelines and the semantic layer, or is the architecture set elsewhere?" This is diligence, and it also surfaces whether the seat can grow.
6. **Anforderungsniveau K.** Ask directly for the band range for this level with Regionalzulage Stufe 1. It is a published, collectively-bargained system — asking is normal, not pushy.
7. **Why leave Swisscom.** Have an answer ready that is about the work (physical infrastructure, asset condition data, a visible public mission) rather than about the pay band, since the pay band is likely lower.
---
## 10. Cover letter
**Not provided — correctly.** SBB states: *"Darum verzichten wir bei dieser Stelle auf ein Motivationsschreiben."* Under the `cl_reference.md` gate a letter is written only when required or when it adds what the resume cannot. Neither holds, and submitting one anyway would signal the posting was not read.
**Does the resume stand alone? Yes.** The lead bullet delivers the operator identity within the first ten seconds, the skills block resolves the stack question, and the Bosch block carries the domain thread without a letter needing to explain it. The one thing a letter could have added — motivation for a lateral move to a public infrastructure operator — is better delivered on the phone call, where it can be a conversation instead of an assertion.
---
## 11. Mechanical verification
- [x] Canonical validator: **PASS**, 0 warnings
- [x] Compiles: **2 pages**, exactly as required
- [x] `pdftotext` order verified: employer → dates → title → location on every entry
- [x] Both pages visually inspected: no clipping, overlap, tiny text, isolated heading or awkward break
- [x] Umlauts (Universität, München) render correctly
- [x] Forbidden-token scan: none of the seven non-canonical JD tools present in the document
- [x] Scope-discipline scan: no full-ownership verb paired with an org-scale object
- [x] Email is `dennis@thiessen.io` per config
- [x] No date of birth, marital status, gender, children, or photo
- [x] Certifications listed once; no duplication
- [x] Bullet lengths 20–31 words, natural variation, none over 35
- [x] No LOC counts, no test counts, no unverified metric
---
## 12. Verdict
| Axis | Result |
|---|---|
| **Hard gate** | PASS |
| **Evidence Fit** | 79/100 — Core, lower end |
| **Document Quality** | 90/100 |
| **Channel Strength** | Weak (upgradeable to Moderate with one phone call) |
**The document is ready.** One Tier 1 fix is worth making (query performance optimization) and two Tier 2 terms are cheap. None of them is blocking.
**The package is not the constraint — the decision is.** Do not submit before the Anforderungsniveau K answer. A 90-quality document against a capped band that lands near current compensation is effort spent to create a decision you have already made twice (BKW, and the standing 180k bar). Make the call first.
output/SBB_DataEngineer_AssetMgmt/e2e_sbb_data_engineer_asset_mgmt_resume.tex
+86
View File
@@ -0,0 +1,86 @@
% SBB — Data Engineer (m/w/d), Asset Management Infrastrukturanlagen (Job ID 103755)
% Profile: International Tech, English, 2 pages. Languages line names German as mother tongue
% per user instruction 2026-08-21.
% Generated from resume_builder/canonical/claims.json + normalized experience files.
% Seven tools named in the JD are non-canonical and are deliberately absent from this
% document; they are listed and bridged in the session file's Gap Assessment. Do not
% reintroduce them here, not even in a comment -- the validator scans raw file text.
\documentclass{resume}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}
\usepackage[a4paper,left=0.65in,right=0.65in,top=0.55in,bottom=0.55in]{geometry}
\usepackage{xcolor}
\usepackage{hyperref}
\hypersetup{hidelinks}
\usepackage[version=4]{mhchem}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[R]{\small \thepage/\pageref{LastPage}}
\name{Dennis Thiessen, M.Eng.}
\headline{Staff Data Engineer $\vert$ Cloud Data Platforms $\vert$ Pipeline Ownership \& Governed Data Products}
\contactline{Bern, Switzerland $\vert$ \href{mailto:dennis@thiessen.io}{dennis@thiessen.io} $\vert$ +41 795 955 585 $\vert$ \href{https://linkedin.com/in/dennis-thiessen}{LinkedIn}}
\begin{document}
\begin{rSection}{Summary}
Staff data engineer who builds and runs production data pipelines and governed data products on AWS at Swisscom, with Component Ownership covering data quality, governance and on-call operation. Earlier work at Bosch centred on semiconductor manufacturing data -- defect records, wafer inspection images and electrical parameter tests -- and on anomaly detection over streamed logs.
\end{rSection}
\begin{rSection}{Technical Skills}
\skillline{Programming}{Python, SQL, PySpark; Java and C\# (professional, historical)}
\skillline{Pipelines and streaming}{Apache Kafka, Apache Airflow, ETL/ELT, data modelling, query performance tuning, Oracle, Teradata, Hadoop/Impala}
\skillline{Cloud and platform}{AWS (S3, Glue, Athena/Iceberg, Redshift, Lambda, Step Functions), CloudFormation, Kubernetes, Docker, GitLab CI/CD, Ansible}
\skillline{Data products and analytics}{Data Mesh data products, active metadata and lineage (Atlassian Compass), data governance; TIBCO Spotfire (co-owned, C\# extensions), Grafana, Prometheus, ELK (proof of concept)}
\skillline{Certifications}{AWS Solutions Architect -- Associate; Data Engineering with AWS; iSAQB CPSA-F; ITIL Foundation}
\skillline{Languages}{German (native), English (fluent)}
\end{rSection}
\begin{rSection}{Professional Experience}
\begin{rSubsection}{Swisscom (Schweiz) AG}{Oct 2023 -- Present}{Staff Data, Analytics \& AI Engineer (promoted from Senior, Apr 2025)}{Bern, Switzerland}
\item Own Fulfillment ETL pipeline components as Component Owner, covering production operation, data quality, governance, incident handling and on-call duty across Oracle, Kafka, Python and Teradata processing.
\item Migrated pipelines in the Fulfillment and Product Analysis domains onto Swisscom's AWS platform using Glue, Athena with Apache Iceberg, Redshift, Airflow and CloudFormation, contributing to the wider company migration programme.
\item Build and model governed data products within Swisscom's company-wide Data Mesh, onboarding source systems and maintaining active metadata and lineage so downstream teams can find and reuse data.
\item Build and operate Python data applications on Kubernetes with GitLab CI/CD in an agile DevOps team, covering automated tests, packaging and deployment from development through production operation.
\item Gather requirements with business stakeholders and deliver data products, analyses and dashboards; automate recurring technical processes and run root-cause analysis under 2nd- and 3rd-level support.
\end{rSubsection}
\end{rSection}
\newpage
\begin{rSection}{Professional Experience (continued)}
\begin{rSubsection}{Robert Bosch Semiconductor Manufacturing Dresden GmbH}{Feb 2020 -- Dec 2022}{Senior Engineer, Data Analysis (Data Engineering)}{Dresden, Germany}
\item Developed Python, Java and C\# data services over Oracle and Hadoop/Impala, tuning query performance for analysis teams working with semiconductor manufacturing data -- defect-management records, wafer inspection images and electrical parameters from process control monitoring.
\item Served as Application Owner for analytics applications and upstream pipelines, defining SLOs, coordinating vendors and delivering user training and documentation.
\item Integrated containerized ML inference with Docker, Kubernetes and Ansible into a continuously operating semiconductor-fab environment, replacing manual image-based defect classification.
\item Built an anomaly-detection proof of concept on Elasticsearch, Logstash, Kibana and Kafka, with Grafana, Prometheus and Loki monitoring for continuously operating manufacturing systems.
\item Co-owned the TIBCO Spotfire environment, built C\# extensions and wafer-map visualizations, and co-presented the work at TIBCO Analytics Forum 2022.
\end{rSubsection}
\begin{rSubsection}{Fraunhofer CML}{Sep 2018 -- Oct 2019}{Research Software Engineer}{Hamburg, Germany}
\item Set up Jenkins CI/CD with quality gates, developed C\#/.NET software for a maritime crew-scheduling system and built containerized microservices for a research data-exchange platform.
\end{rSubsection}
\begin{rSubsection}{Vizrt}{Jul 2017 -- May 2018}{Test Automation / DevOps Engineer}{Bergen, Norway}
\item Developed automated audio/video integration and unit tests in Python and connected them as quality gates in the CI/CD pipeline.
\end{rSubsection}
\begin{rSubsection}{Generali Deutschland Informatik Services GmbH}{May 2015 -- Jun 2017}{IT Consultant}{Hamburg, Germany}
\item Introduced BDD test automation through a proof of concept, held technical responsibility for the suite and Jenkins jobs, and trained colleagues in the Java community.
\end{rSubsection}
\end{rSection}
\begin{rSection}{Education}
\compactentry{M.Eng. Computer Aided Engineering (Software Design \& Engineering), Universität der Bundeswehr München}{Apr 2012 -- Oct 2013}
Thesis at Tongji University, Shanghai: \textit{Development of a Web-Based Remote Fault Diagnosis System}; grade 1.0.
\compactentry{B.Eng. Information and Telecommunication Technologies, Universität der Bundeswehr München}{Oct 2009 -- Oct 2012}
\end{rSection}
\end{document}
output/SBB_DataEngineer_AssetMgmt/interview_brief_sbb_2026-09-09.md
+111
View File
@@ -0,0 +1,111 @@
# Interview Brief — SBB, Data Engineer Asset Management Infrastrukturanlagen
**When:** Mittwoch, 9. September 2026, 08:30 · 45 Minuten · Microsoft Teams
**Stage:** **2 of 4.** The posting's own process is: Bewerbung → **Virtuelles Kennenlernen mit HR und Führungskraft** → Persönliches Kennenlernen → Finaler Entscheid.
**Who is likely on the call:** HR **plus Andri Wienandts, People Leader** (+41 79 364 62 53). This is *not* a pure HR screen — the hiring manager is in the room.
**Job ID:** 103755 · Bern / Work Smart · 60–100% · **Anforderungsniveau K, Regionalzulage Stufe 1**
**Submitted:** 2026-08-25 · Core · Evidence Fit 79/100 · Document Quality 92/100 · **cold channel**
> **45 minutes shared between two people is short.** Assume ~25 min of them asking, ~10 min of you asking, ~10 min logistics. You will not get a second cheap shot at the money and scope questions before the in-person round, so do not let them fall off the end.
---
## 1. The one thing to understand before you dial in
**A cold submit converted.** You never called Andri Wienandts, despite his mobile being printed on the posting — the channel plan said to call first and it did not happen. It got an interview anyway.
More telling: your resume **never contains the words Snowflake, dbt or Power BI**, and all three are named in the required-qualifications block. An ATS keyword screen configured on those literals would have dropped you. It didn't. **Someone read the dossier and made a judgement.** That means the honest substitution story (Teradata/Oracle/AWS → Snowflake; Airflow/PySpark → dbt; Spotfire → Power BI) already survived its first contact with a human — you are not walking in needing to apologise for it. Defend it the same way, calmly, as a tooling difference rather than a gap.
This is also the **first interview in the evidence-first cohort** — 6 applications, 2 rejections, this is the first conversion.
---
## 2. Your genuine edge — lead with this
For a role whose stated purpose includes **vorausschauende Wartung** on physical infrastructure assets, you have a combination most candidates on this req will not:
- **Industrial sensor and asset data, in production.** Bosch fab: defect-management records, wafer inspection images, PCM electrical parameters (BS-2).
- **Anomaly detection on streaming data.** The ELK/Kafka PoC with Grafana/Prometheus/Loki monitoring (BS-4).
- **Containerised ML inference into a plant that cannot stop.** Docker/Kubernetes/Ansible into 24/7 semiconductor production, after which defect-image classification ran fully automated (BS-1).
- **A master's thesis on vibration-based condition monitoring of machine tools** — rule-based reasoning plus a 7-10-3 ANN.
- **The operational half of this job, today, at a peer Swiss enterprise in the same city.** Component Owner with on-call, data quality, governance and SLA accountability (SW-2) is precisely *"Betrieb"* and *"Verantwortung für die Qualität und Nachhaltigkeit technischer Lösungen"*.
**Thesis caveat — hold this line:** it is a **methods prototype**, never a validated system. No real operational data (training data came from a cited thesis; test data was modified plus a random generator), **no accuracy figures exist**, and PSO was *surveyed only*, never implemented. The throughput figure (~500 samples/s against 72.9 kSPS sensors) is barred from external use. Describe it as genuine interest and method familiarity, not as a result.
**Where they beat you:** named-product fluency on the exact stack, and rail domain — the asset taxonomy, EN 50126 RAMS vocabulary, the swissTAMP/RIS ecosystem.
---
## 3. The four questions you must not leave without asking
These were never resolved before submitting. They are now live screening topics, and this call is the natural place for them.
### Q1 — Anforderungsniveau K, the actual band
> *"Die Stelle ist im Anforderungsniveau K mit Regionalzulage Stufe 1 ausgeschrieben. Könnt ihr mir die Bandbreite für dieses Niveau nennen, und wo eine Person mit meinem Profil typischerweise einsteigt?"*
**Why it is fair game:** SBB publishes its pay system and campaigns publicly on Lohngleichheit and transparency. Asking about a collectively-bargained band is normal, not pushy. The figures are simply not public, so asking is the *only* way to get them.
**Your bar:** Swisscom ~140k; you move at 180k+ all-in. **K is very likely below that.** The Bern/Thun WLB-exception tier deliberately accepts below-bar comp for a local seat with a real mission — but it has never accepted a below-*level* move (see Q2). Know before the call which trade you are actually willing to make, so you can react in the moment rather than going quiet.
### Q2 — Level and design ownership *(the crux)*
> *"Die Ausschreibung liest sich als Senior-IC-Rolle. Trägt sie Design-Verantwortung für die Pipelines und das semantische Modell, oder wird die Architektur an anderer Stelle gesetzt?"*
**Why this decides it.** Every other application in your log was a stretch *upward*. This one reads *below* your current Staff (Engineer IV) + Component Owner scope: it asks you to *mentor* other data engineers but the only architecture language is *"Datenarchitekturen entwerfen"* as a personal skill, not an ownership mandate. **This is the same shape as BKW, which you declined as a lateral in this same Bern tier.** Ask it as diligence — it also surfaces whether the seat can grow.
### Q3 — RAMSI, and how much of the job it really is
> *"Was ist RAMSI genau, und welchen Anteil der Rolle macht die Weiterentwicklung der Webanwendung aus?"*
One of six responsibility lines is web development in **Java Spring Boot / Angular**. Your Java is professional but historical, Spring Boot was a PoC contribution roughly a decade ago, and **Angular is absent entirely**. Do not oversell it — ask what share it is. It may be work you do not want, and that is worth knowing now.
**Do not assert what RAMSI is.** It is not publicly documented; the RAMS/EN 50126 family is an *inference*. Ask.
### Q4 — Kidz Care, if childcare matters to the decision
SBB covers *"bis zu 90 %"* of Betreuungskosten but scales it on **gross household income**, and the bracket table is intranet-only. At this band, realistically far below 90%. Worth one question if it affects the maths.
---
## 4. Questions they will ask — and the honest answer
| Their question | Your line |
|---|---|
| **Snowflake? dbt?** | "Not by name. I have run the same shape of work on Teradata and Oracle, then on AWS — Glue, Athena with Iceberg, Redshift — with Airflow for orchestration. The modelling, testing and lineage discipline transfers; the syntax is what I'd pick up." |
| **Power BI?** | "I co-owned a BI platform rather than just consuming one — TIBCO Spotfire at Bosch, including C# extensions and wafer-map visualisations, which I co-presented at the TIBCO Analytics Forum 2022." |
| **Argo Workflows / Helm?** | Airflow is a direct orchestrator substitution; Kubernetes is current and production. **Do not claim Helm** — it is not canonical. |
| **Predictive maintenance experience?** | Lead with Bosch sensor/process data and the ELK/Kafka anomaly-detection PoC. Then the thesis, framed as a methods prototype. |
| **Why leave Swisscom?** | **Make it about the work, not the band.** Physical infrastructure, asset condition data with a visible public consequence, a Bern seat, and a technical brief that sits on your actual thesis. Do not volunteer that you expect a pay cut. |
| **Trade-offs between latency, consistency and cost?** | A named required skill and a real strength — Component Owner for business-critical Oracle/Kafka → Teradata ETL with on-call and SLAs. Have one concrete example where you chose batch over streaming, or accepted staleness for cost. |
| **Mentoring / fachliche Entwicklung?** | Application Owner at Bosch: documentation, user training, vendor coordination; Swisscom Security Champion for the team. Real, but framed as team roles rather than line management. |
**Two things to keep off the table:** don't claim Snowflake, dbt, Power BI, Helm, Argo, Spring Boot or Angular by name; and don't quote any unverified metric — nearly everything in your record is `metrics: unverified`, and inventing a number in an interview is the same error as putting one on the resume.
---
## 5. Company context worth showing you know
- **swissTAMP** is SBB's central source for asset condition and behaviour data — inspection, measurement and operating data, trends and forecasts. SBB describes it explicitly as enabling the move *"from reactive through proactive to predictive maintenance."*
- Alongside it: **RIS** (Räumliche Informationssysteme, incl. Datenbank feste Anlagen) and **Mess- und Diagnosetechnik (MUD)**, which runs automated condition monitoring of track, switches and power supply.
- SBB operates a **Competence Center Predictive Maintenance**.
- **Funded and growing, not a backwater:** there is an active EU-journal tender (235005-2026) for up to **three new-generation infrastructure diagnostic vehicles**. More diagnostic vehicles means more sensor volume — which is exactly what this pipeline seat exists to absorb. Good, specific thing to reference.
- **Culture:** `#einfachDu` — informal *du* from first contact, so match that register. Work Smart hybrid. Cover letter deliberately waived. 60–100% explicitly welcome.
---
## 6. Practicalities
- **08:30 start** — be on Teams early; test audio the night before.
- Match the **du** register from the first sentence.
- The posting is in German and the team is German-speaking; **your resume was submitted in English** with a Languages line naming German as mother tongue. Expect the call in German. Don't be thrown if they open in English.
- Bern-based, German native, EU citizen with a Swiss B permit, no relocation, no sponsorship — **the cleanest practical fit in your whole log.** Say so if the topic opens.
- Have ready: a concrete latency/consistency/cost trade-off story, and one example of mentoring or lifting another engineer.
---
## 7. What to decide *before* 08:30
Not homework for them — homework for you:
1. **If K comes in at, say, 130–150k all-in: is this still a yes?** The WLB-exception tier says a Bern seat with a real mission can go below the 180k bar. It has never said below-level scope is acceptable.
2. **If Q2 comes back "architecture is set elsewhere": is it still a yes?** That is the BKW shape you already declined once.
3. **If RAMSI turns out to be 40% Angular work: is it still a yes?**
Any one of those can be a no, and it is much better to know your own answer before they say the number than to improvise it on camera at 08:40.
output/SBB_DataEngineer_AssetMgmt/resume.cls
+59
View File
@@ -0,0 +1,59 @@
\ProvidesClass{resume}[2026/07/27 Evidence-first professional resume]
\LoadClass[11pt,a4paper]{article}
\RequirePackage{enumitem}
\RequirePackage{ifthen}
\RequirePackage{lastpage}
\RequirePackage{parskip}
\RequirePackage{needspace}
\pagestyle{empty}
\setlength{\parindent}{0pt}
\def\@resumename{}
\def\@resumeheadline{}
\def\@resumecontact{}
\newcommand{\name}[1]{\def\@resumename{#1}}
\newcommand{\headline}[1]{\def\@resumeheadline{#1}}
\newcommand{\contactline}[1]{\def\@resumecontact{#1}}
\newcommand{\printresumeheader}{%
{\LARGE\bfseries \@resumename}\par
\vspace{0.15em}
\ifthenelse{\equal{\@resumeheadline}{}}{}{\@resumeheadline\par}
\vspace{0.1em}
{\small \@resumecontact}\par
\vspace{0.45em}
}
\AtBeginDocument{\printresumeheader}
\newenvironment{rSection}[1]{%
\vspace{0.35em}
{\bfseries #1}\par
\vspace{0.1em}\hrule\vspace{0.35em}
}{%
\vspace{0.2em}
}
% Arguments: employer, dates, formal title, location.
\newenvironment{rSubsection}[4]{%
\Needspace{6\baselineskip}%
{\bfseries #1}\hfill{\small #2}\par
{\itshape #3}\hfill{\itshape #4}\par
\vspace{0.1em}
\begin{itemize}[leftmargin=1.25em,label=\textbullet,itemsep=0.18em,topsep=0.15em,parsep=0pt,partopsep=0pt]
}{%
\end{itemize}
\vspace{0.25em}
}
\newcommand{\skillline}[2]{%
\textbf{#1:} #2\par
\vspace{0.08em}
}
\newcommand{\compactentry}[2]{%
\textbf{#1}\hfill #2\par
}
output/SBB_DataEngineer_AssetMgmt/session_sbb_data_engineer_asset_mgmt.md
+307
View File
@@ -0,0 +1,307 @@
# Session: SBB (Schweizerische Bundesbahnen) — Data Engineer (m/w/d), Asset Management Infrastrukturanlagen
## STATUS: INTERVIEW INVITED 2026-08-28 (submitted 2026-08-25)
**Interview: Mittwoch 9. September 2026, 08:30, 45 Minuten, Microsoft Teams.**
Per the posting's own process this is **stage 2 of 4** - *"Virtuelles Kennenlernen mit HR und Fuehrungskraft"* - so **HR together with Andri Wienandts**, not a pure HR screen. Brief written to `interview_brief_sbb_2026-09-09.md`.
**Two facts worth recording.** (1) This is the **first interview of the evidence-first cohort** - 6 applications, 2 rejections, 1 conversion. (2) It converted from a **cold submit**: the channel plan called for phoning Andri Wienandts before applying and that never happened, so the published warm contact was never used. The resume also never contained **Snowflake, dbt or Power BI**, the three literals an ATS keyword screen would have keyed on - the session recorded that risk as "accepted, not solvable". It did not filter him out, which means a human read the dossier and the honest-substitution framing survived first contact. Do not over-read one data point, but it is evidence against the assumption that the honest omissions are fatal at the screen.
**Comp worksheet:** `comp_worksheet_sbb.md` - break-even framework to complete **before** the call, so a named K figure can be judged on the spot. Key points recorded there: base-to-base is meaningless (free GA, Regionalzulage and the employer pension share are real money invisible in a base comparison, while the Swisscom bonus is real money SBB will not match); the GA is a taxable Lohnausweis benefit so its net value is below list; and the commute saving and the GA are the same franken - count once. **K is a capped GAV band with no negotiation lever at the top**, so this is a yes/no against a fixed number. **User's own read, 2026-08-28: he does not expect it to clear.** He is taking the call anyway for the questions, for interview practice with NATO JWC / BIS / Microsoft Principal still live, and because **the K figure is the first real data point on what the whole Bern/Thun local tier pays** - RUAG, Swissgrid, BKW and BFH sit in comparable band structures. Record the figure afterwards regardless of outcome.
**Still unresolved, and now live call topics:** Anforderungsniveau K band, design/architecture scope vs current Staff + Component Owner level, the RAMSI Java/Spring Boot/Angular share of the role, and the Kidz Care actual rate at this band.
Application sent by the user on 2026-08-25 (2pp English resume, no cover letter — SBB waives it;
the Tier 1 query-performance fix was already applied on 2026-08-21, so the submitted document is
the 92/100 version). Cohort slot consumed: **Core 2/7** in `evidence-first-2026-01`.
**Submitted with the comp question still open.** The channel plan below called for phoning Andri
Wienandts *before* applying, and the critique's closing line advised against submitting ahead of the
Anforderungsniveau K answer. That did not gate the submission. Consequence, carried forward: the K
band and the design-scope question are now **screening-call topics, not pre-cleared facts**, and the
same is true of Kidz Care childcare support (SBB covers "bis zu 90 %" of Betreuungskosten but scales
it on *gross household income*, and the bracket table is intranet-only — assume far below 90 % at
this band). Prep an honest answer before any call rather than discovering the number in it.
## JD Integrity
- File/source: `JD_SBB_DataEngineer_AssetMgmt.txt` (this folder) — copied from `JDs/JD_SBB_DataEngineer_AssetMgmt.txt`
- Source URL: https://jobs.sbb.ch/v2/offene-stellen/data-engineer-m-w-d-asset-management-infrastrukturanlagen/45fa2411-31dc-4520-827e-1c93a527534c
- Retrieval method and date: Playwright headless scrape via `job_scout/.venv`, 2026-08-21 (board is a JS-rendered SPA; WebFetch is JS-blind on it)
- Verbatim posting: YES — full visible posting body, German original, no paraphrase
- Posting status: LIVE as of 2026-08-21. Discovered on the 2026-08-21 scout run, posted 2026-08-20. Job ID 103755.
- Named contact on posting: Andri Wienandts, People Leader, +41 79 364 62 53
- Notable: SBB explicitly waives the cover letter for this req ("verzichten wir bei dieser Stelle auf ein Motivationsschreiben")
## Application Decision
- Audience profile: **International Tech, in English** — user decision 2026-08-21. Phase 0 had provisionally recommended Swiss/DACH on the strength of the German-language posting and `#einfachDu` register; the user chose English, and since the document carries no photo and no demographic data, International Tech (the config default) is the coherent profile. The single Swiss/DACH element retained is an explicit **Languages line naming German as mother tongue**, at the user's instruction — it answers the obvious recruiter question on a German-language posting.
- Evidence Fit: **79/100**
- Fit class: **Core** (75+, no hard-gate failure)
- Hard gate: **PASS** — no minimum qualification is a Gap; the title-defining capability (build/operate cloud data pipelines and data products) is Direct and current; no clearance, relocation, language or authorization issue.
- Channel Strength: **Moderate (achievable)** — the posting names the People Leader with a direct mobile number. Every application in the log to date has been Weak/cold; this is the first with a published warm entry point. Rated Weak until the call actually happens.
- Channel plan: call Andri Wienandts **before** applying. Two questions to settle: (1) the Anforderungsniveau K salary band ceiling, (2) whether the seat carries architecture/design scope or is purely build-and-run.
- Cohort slot: Core would be **2/7** (Google SWE III consumed 1/7). Adjacent 1/2, Stretch 0/1 unchanged.
- Decision: **PROCEED to Phase 1 — but only after the comp question is answered.** The technical and practical fit is real and the hard gate passes cleanly. The blocker is not fit: it is that this seat is a lateral-or-below step in a capped federal pay band. If Anforderungsniveau K tops out near the current 140k, Phase 1 is wasted work no matter how good the framing is.
### Evidence Fit breakdown (per `critique_framework.md`)
| Dimension | Weight | Score | Reasoning |
|---|---:|---:|---|
| Required qualifications | 35 | **29** | 6 required lines. R1, R2, R3 (core), R6 Direct; R4 Direct–Adjacent; R5 split — SQL/modelling Direct, Power BI a named Gap. |
| Core responsibilities | 25 | **20** | P1, P2 (integration), P3 (data products), P5, P6 Direct; P2 predictive maintenance Adjacent-strong; **P4 (Spring Boot/Angular) a Gap**. |
| Level and ownership | 15 | **9** | The dimension asks whether title, decisions and leadership *match*. They do not: the seat sits **below** current Staff + Component Owner scope. Capability is not the issue — fit of level is. Revealed behaviour already rejected this exact shape once (BKW, same Bern tier, declined as a lateral), so scoring it as a near-match would contradict the Competitive Read below. |
| Recency and depth | 10 | **10** | The strongest evidence is his current job, at a peer Swiss enterprise, today. Nothing here is historical. |
| Domain/tool transfer | 10 | **6** | Snowflake, dbt, Argo, Helm, Power BI, Spring Boot, Angular all non-canonical. Each is individually substitutable and the JD says "z. B." — but seven named tools is a real deduction. |
| Practical constraints | 5 | **5** | Bern resident, German native, EU citizen + B permit, hybrid, no clearance. Best in the log. |
| **Total** | **100** | **79** | **Core** (75+), at the lower end. For calibration: Google SWE III 89, Kraken 87.2, MS ISE 85.8, MS Principal FDE 84.2, BIS 80.5, Aker BP 79, NATO JWC 77.5. Reads as *good fit with one unresolved blocker*, not *best fit in the log*. |
## ATS Keywords
Extracted from the JD. **Bold = canonical and safe to use.** Struck-through = named in the JD but **not canonical — must not appear in any output document.**
- **Core function:** **Data Engineer**, **Datenpipelines**, **ETL/ELT**, **Data Warehouse**, **Batch**, **Streaming**, **Orchestrierung**, **Datenprodukte**, Enterprise Analytics Plattform, semantische Modelle
- **Platform/cloud:** **cloudbasierte Datenplattformen**, **Kubernetes**, **Docker**, **AWS**, **CI/CD**, **GitLab CI/CD**, ~~Helm~~, ~~Argo Workflows~~
- **Data stack:** **Kafka**, **Airflow**, **PySpark**, **SQL**, **Datenmodellierung**, Star-/Snowflake-Schemata, **APIs**, ~~Snowflake~~, ~~dbt~~
- **Languages/SWE:** **Python**, **Tests**, **Packaging**, **Deployment**, **Java** (with context), ~~Spring Boot~~, ~~Angular~~
- **BI/reporting:** **Reporting**, **Dashboards**, **Abfrageleistungsoptimierung**, TIBCO Spotfire (canonical, as the bridge), ~~Power BI~~
- **Domain:** **Anlagenmanagement**, **Infrastrukturanlagen**, **vorausschauende Wartung / Predictive Maintenance**, **Sensordaten**, **Anomalieerkennung**, Zustandsdaten, RAMS, RAMSI
- **Ways of working:** **agil**, **Betrieb**, **Datenqualität**, **Governance**, **On-Call**, **Verantwortung**, **Mentoring / fachliche Entwicklung**, **Stakeholder**, Fachspezialist:innen
**ATS risk, stated plainly:** an automated screen keyed on the literals *Snowflake*, *dbt* or *Power BI* would reject this resume, because none of the three can honestly appear. That risk is accepted, not solvable — it is exactly the case the warm channel exists to bypass.
## Gap Assessment
| Gap | Severity | Why it is not a hard gate | How it is handled |
|---|---|---|---|
| **Snowflake** | Medium — named in a required line *and* a responsibility | Prefixed "z. B."; the requirement is cloud data platforms with a modern stack | Bridge from Teradata/Oracle → AWS warehouse work (SW-1) and data products (SW-7). Never write the word. |
| **Power BI** | Medium — named in a required line | BI tooling is genuinely substitutable; he **co-owned** a BI platform | Bridge from TIBCO Spotfire co-ownership + C# extensions + TAF 2022 (BS-5). |
| **dbt** | Low–Medium | "z. B." | Bridge from Airflow + PySpark transformation/testing discipline. |
| **Argo Workflows** | Low | "z. B."; direct orchestrator substitution | Bridge from Airflow. |
| **Helm** | Low | Not a standalone requirement; implied by Kubernetes | Kubernetes is canonical; **Helm is not — do not list it.** |
| **Spring Boot / Angular (RAMSI)** | Medium — a whole responsibility line | One of six responsibilities; not the title, not in the required list | Java is `allowed-with-context` (GN-3, BS-2). Do not claim Spring Boot or Angular. Raise it honestly in the call — he may not want this work. |
| **Rail/RAMS domain** | Low | No JD line requires it | Bridge: industrial asset/sensor data at Bosch; condition-monitoring thesis. Ask about RAMSI rather than assuming. |
| **Anforderungsniveau K comp** | **Unresolved — decision-blocking for the user, not a fit gap** | Not a qualification at all | Ask the People Leader. Do not research further; the figures are not public. |
**No Gap sits on a minimum qualification, and the title-defining capability is Direct.** Hard gate: **PASS.**
## Requirements
Requirement text is the JD's German original, abbreviated. "Required" = listed under *Das bringst du mit*; "Responsibility" = listed under *Das kannst du bewegen*.
| # | Requirement | Required/preferred | Direct/Adjacent/Gap/Constraint | Canonical evidence | Gate? |
|---|---|---|---|---|---|
| R1 | Kommunikation, Teamorientierung, **fachliche Entwicklung anderer Dateningenieur:innen fördern** | Required | **Direct** | Staff Engineer (Engineer IV) since 2025-04; BS-3 Application Owner incl. training and documentation; GN-1 training ownership | No |
| R2 | Agile, selbständige Arbeitsweise; Verantwortung für Qualität und Nachhaltigkeit technischer Lösungen | Required | **Direct** | SW-2 Component Owner — operation, data quality, governance, incidents, on-call | No |
| R3 | Mehrjährige Erfahrung in Entwurf, Aufbau, Entwicklung und **Betrieb cloudbasierter Datenplattformen** mit modernem Data Stack (ETL/ELT, DWH, Streaming, Orchestrierung — **z. B.** Snowflake, dbt, Helm, Argo Workflows, Kafka) | Required | **Direct** (core) / **Gap** (named products) | SW-1 AWS migration of his domains' ETL from Teradata/Oracle; SW-2; SW-3 Python apps on Kubernetes + GitLab CI/CD; SW-7 governed data products; skills: Kafka, Airflow, AWS, PySpark, Kubernetes all production-current | **No** — the product list is prefixed **"z. B."** (for example), so it enumerates a stack archetype, not mandates. See Gap note below. |
| R4 | Skalierbare und **kosteneffiziente Datenarchitekturen** entwerfen; Trade-offs Latenz / Konsistenz / Kosten beurteilen | Required | **Direct–Adjacent** | SW-1, SW-2, SW-3. Design/trade-off reasoning is real and current; **cost-efficiency figures are `metrics: unverified`** — describe the practice, never quantify a saving | No |
| R5 | **Datenmodellierung (Star-/Snowflake-Schemata)**, Reporting mit **Power BI**, fortgeschrittenes SQL, Abfrageleistungsoptimierung | Required | SQL **Direct**; modelling **Direct**; dimensional schemas **Adjacent**; Power BI **Gap (Adjacent via Spotfire)** | SQL production-current; SW-7 models data products (allowed verb "model"); BS-5 co-owned TIBCO Spotfire environment and built C# extensions, co-presented at TIBCO Analytics Forum 2022 | **No** — BI tooling is genuinely substitutable and he co-owned a BI platform. But this is the single weakest required line. |
| R6 | **Software-Engineering in Python** — Tests, CI/CD, Packaging, Deployment Best Practices | Required | **Direct (strongest line in the JD)** | Python production-current; SW-3 GitLab CI/CD on Kubernetes; VZ-2 automated tests wired to CI/CD quality gates; GN-1 BDD + Jenkins | No |
| P1 | Datenpipelines auf der **Enterprise Analytics Plattform** aufbauen, **betreiben**, weiterentwickeln (Batch **und** Streaming) | Responsibility | **Direct** | SW-1, SW-2, SW-3; Kafka + Airflow + PySpark production-current. "Betreiben" maps exactly to Component Ownership incl. on-call. | No |
| P2 | Datenquellen via **APIs/Kafka** integrieren; ermöglicht Analysen, Dashboards und **vorausschauende Wartung** | Responsibility | Integration **Direct**; predictive maintenance **Adjacent (strong)** | Kafka production-current; SW-7 source onboarding. PdM cluster: BS-2 fab sensor/process data (defect records, wafer inspection images, PCM electrical parameters); BS-4 ELK/Kafka anomaly-detection PoC; BS-1 containerized ML inference in 24/7 production; EDU-MENG thesis on vibration-based condition monitoring | No |
| P3 | **Datenprodukte und semantische Modelle** auf **Snowflake und Power BI** gestalten | Responsibility | Data products **Direct**; both named products **Gap** | SW-7 — builds governed data products and onboards sources inside Swisscom's company-wide Data Mesh, Atlassian Compass for metadata/lineage | No |
| P4 | **RAMSI** weiterentwickeln — Webanwendung (**Java Spring Boot / Angular**), Reports, Dashboards | Responsibility | **Gap** — the clearest one | Java is `production-historical`, `allowed-with-context` (GN-3 Java/J2EE workflow features; BS-2 Java data services). **Spring Boot: absent from canonical. Angular: absent; JavaScript is `production-historical-limited`.** | **No** (one of six responsibilities, not the title) — but see Open Risks |
| P5 | Anforderungen mit dem Business aufnehmen; nachhaltige Lösungen im Betrieb | Responsibility | **Direct** | SW-4 delivers data products/dashboards/analyses with internal B2B stakeholders and product owners | No |
| P6 | Wissen im Team teilen, zur fachlichen Weiterentwicklung beitragen | Responsibility | **Direct** | Same as R1 | No |
| C1 | Standort **Bern** / Work Smart (hybrid), 60–100% | Constraint | **Direct — best in the entire application log** | Lives in Bern. No relocation, no commute problem, hybrid matches the stated 2–3 day maximum | No |
| C2 | German-language posting and process | Constraint | **Direct** | German native | No |
| C3 | Work authorization | Constraint | **Direct** | German citizen (EU), Swiss B permit, no sponsorship required | No |
| C4 | **Anforderungsniveau K, Regionalzulage Stufe 1** | Constraint | **UNRESOLVED — the real risk** | SBB GAV runs 15 collectively-bargained levels A–O with a fixed Basiswert/Höchstwert per level; K is the 11th. Current K figures are **not published** in the sources reachable on 2026-08-21 (the SEV brochure documents the mechanism, not the table; the GAV PDF download returns an error page). | **Not a fit gate — a user-decision gate.** See Open Risks. |
### Gap note on R3 (important — do not overstate it)
Five products are named: Snowflake, dbt, Helm, Argo Workflows, Kafka. Only **Kafka** is canonical. Snowflake, dbt, Argo Workflows and Helm are **absent from `claims.json`** — under `shared_ops.md` ("omit anything absent, ambiguous or unverified") none of them may appear in any output document.
This is survivable **only because the JD writes "z. B."**. The requirement as written is *cloud data platforms with a modern data stack*, exemplified by those tools. The substitution map is honest and defensible in an interview:
| JD names | Canonical equivalent | Transfer honesty |
|---|---|---|
| Snowflake | Teradata / Oracle / AWS warehouse work (SW-1), data products (SW-7) | Strong — cloud MPP warehouse concepts, but **never claim Snowflake itself** |
| dbt | Airflow + PySpark transformation pipelines | Good — same modelling/testing/lineage discipline, different tool |
| Argo Workflows | Airflow | Strong — direct orchestrator substitution |
| Helm | Kubernetes (production-current-and-historical) | Adequate — **Helm itself is not canonical; do not list it** |
| Power BI | TIBCO Spotfire co-ownership (BS-5) | Good — BI platform ownership transfers; the product does not |
## Competitive Read
- **Obvious-fit candidate:** a Swiss data engineer with 5+ years on Snowflake + dbt + Power BI, ideally already inside a Swiss infrastructure/utility/transport operator, German-speaking, and comfortable maintaining a Java/Angular internal web application. That last trait is the unusual one — most pure data engineers do not want it, which measurably widens the field for Dennis.
- **Dennis's advantage:** he does the *operational* half of this job today, at a peer Swiss enterprise, in the same city. Component Ownership with on-call, data quality and governance is exactly "Betrieb" and "Verantwortung für Qualität und Nachhaltigkeit". Add the genuinely rare combination for this specific req: **industrial sensor/asset data (Bosch fab: defect records, wafer inspection images, PCM parameters) + anomaly detection on streaming data (ELK/Kafka) + a master's thesis on vibration-based condition monitoring of machine tools.** For a role whose stated purpose includes *vorausschauende Wartung* on physical infrastructure assets, that combination is worth more than Snowflake syntax.
- **Their advantage:** named-product fluency on the exact stack (Snowflake, dbt, Power BI) and, for the strongest candidates, rail domain knowledge — the asset taxonomy, EN 50126 RAMS vocabulary, swissTAMP/RIS ecosystem.
- **Level/scope comparison:** **This is the inversion, and it is the crux of the application.** Every other role in the log was a level stretch upward; this one reads *below* current scope. Dennis is Staff (Engineer IV) with Component Ownership at Swisscom; this is a senior IC data-engineer seat that asks him to *mentor* other data engineers but not to own architecture. The JD's only architecture language is "Datenarchitekturen entwerfen" as a personal skill, not an ownership mandate. Same shape as BKW, which the user declined as a lateral in this same Bern WLB-exception tier.
## Company Context
- **Employer:** Schweizerische Bundesbahnen SBB, division Infrastruktur — Anlagenmanagement (infrastructure asset management). Swiss federal railway; Bern is the corporate seat.
- **Where the role sits:** SBB runs a real, funded asset-management data ecosystem. **swissTAMP** (swiss Track Analysis & Maintenance Planning) is its central source for asset condition and behaviour — inspection data, measurement and operating data, trends and forecasts — and SBB describes it explicitly as enabling the move *"from reactive through proactive to predictive maintenance."* Alongside it sit **RIS** (Räumliche Informationssysteme, incl. Datenbank feste Anlagen) and the **Mess- und Diagnosetechnik (MUD)** department, which does automated condition monitoring of track, switches and power-supply systems across the standard-gauge network. SBB also operates a **Competence Center Predictive Maintenance**.
- **Current investment signal:** SBB has an active EU-journal tender (235005-2026) for up to **three new-generation infrastructure diagnostic vehicles** to sustain and expand condition-data monitoring. More diagnostic vehicles means more sensor volume, which is precisely what an Enterprise Analytics Platform pipeline role exists to absorb. The domain is funded and growing, not a maintenance backwater.
- **RAMSI:** not publicly documented. Almost certainly an internal application in the RAMS family (Reliability, Availability, Maintainability, Safety — EN 50126, the standard railway discipline) applied to Infrastruktur. **Treat as inference — do not state it as known fact in any document or interview; ask about it instead.** Good question for the People Leader call.
- **Culture signals:** `#einfachDu` (mandated informal "du" from first contact), Work Smart flexible/hybrid working, published pay system and active pay-equity commitment, cover letter deliberately waived, 60–100% part-time explicitly welcome. Process: application → virtual meeting with HR + line manager → in-person meeting → decision.
- **"Why them" angle:** public infrastructure whose data work has a physical consequence — keeping the Swiss network running — combined with a Bern seat and a genuinely on-thesis technical brief. This is the first role in the log where the mission and the commute both work without compromise.
## Framing Strategy
- **Professional identity:** Staff-level data engineer who *builds and runs* governed data products on a cloud platform in a large Swiss enterprise — pipeline construction, operation, data quality and stakeholder delivery — with unusual depth in **industrial sensor and asset data**.
- **Lead narrative:** "I already do the operational half of this job at Swisscom, on the same kind of platform, in the same city — and I have handled machine and sensor data since Bosch."
- **Strongest proof points:**
1. SW-2 — Component Owner for business-critical Fulfillment ETL: operation, data quality, governance, incidents, on-call. Maps to "Betrieb" + "Verantwortung für Qualität und Nachhaltigkeit".
2. SW-1 — migrated his domains' ETL from Teradata/Oracle to AWS (scoped, hedged per KB correction: **his domains**, contributor to the wider programme).
3. SW-7 — builds governed data products and onboards sources within Swisscom's company-wide Data Mesh; Atlassian Compass for metadata and lineage. Maps to "Datenprodukte und semantische Modelle".
4. SW-3 + VZ-2 + GN-1 — Python on Kubernetes with GitLab CI/CD, automated tests wired to quality gates. Covers R6 completely.
5. BS-2 / BS-4 / BS-1 — fab sensor and process data, ELK/Kafka anomaly detection, containerized ML inference in a 24/7 production environment. This is the predictive-maintenance bridge and the real differentiator.
- **Honest adjacent bridges (state as transfer, never as the named tool):** Spotfire → Power BI; Airflow → Argo Workflows; Teradata/Oracle/AWS warehouse → Snowflake; PySpark/Airflow transformation discipline → dbt.
- **Explicit gaps (never paper over):** Snowflake, dbt, Argo Workflows, Helm, Power BI, Spring Boot, Angular — none are canonical, none may appear in output. Rail domain knowledge is absent.
- **Emphasize:** operation and ownership over greenfield build; sensor/asset data lineage from Bosch to now; German-language Swiss enterprise delivery; mentoring (R1) since it is a *required* line and Staff level evidences it.
- **Downplay:** GenAI/LLM work entirely — SW-8 is irrelevant here and its forbidden list is long. Also downplay FDE/customer-facing framing carried over from recent applications; this is an internal-stakeholder role (SW-4 is the right frame).
- **Do NOT include:** Security Champion (SW-5 — JD has no security requirement; default is OMIT). LOC/test counts. Any cost or performance metric (`metrics: unverified` across the board).
- **User directives:** none given for this JD.
- **Scope discipline watch (recurring error):** SBB will read "Data Mesh", "data platform" and "migration" as org-scale objects. Never pair a full-ownership verb with one. Correct: "builds governed data products within Swisscom's company-wide Data Mesh", "migrated my domains' ETL stack". Never: "built a Data Mesh", "migrated the warehouse".
## Critique Context
- **Reviewer persona:** a Swiss enterprise engineering lead (Andri Wienandts, People Leader) plus an SBB HR screener. Reads German. Screens for: does this person actually *operate* platforms or only build them; will they stay in a banded federal role; can they talk to Fachspezialist:innen.
- **Competitive landscape:** see Competitive Read. The Java/Angular half of the role narrows the field.
- **Domain vocabulary:** Anlagenmanagement, Infrastrukturanlagen, vorausschauende Wartung / Predictive Maintenance, Zustandsdaten, Diagnose, Enterprise Analytics Plattform, Datenprodukte, semantische Modelle, Batch und Streaming, Orchestrierung, Fachspezialist:innen, RAMS.
- **Likely first technical challenge:** "You have not used Snowflake or dbt — walk me through how your Teradata-to-AWS work maps." Prepare the substitution map above as a spoken answer, and answer the Power BI question with the Spotfire co-ownership honestly.
## Cover Letter Decision
- **NO** (provisional — revisit only if the People Leader call suggests otherwise)
- **Reason:** SBB explicitly waives it — *"Darum verzichten wir bei dieser Stelle auf ein Motivationsschreiben."* Under `cl_reference.md` a letter is generated only when required or when it adds information the resume cannot show. Neither holds. Writing one anyway risks reading as not having read the posting.
- **Better use of the same effort:** the phone call to Andri Wienandts, which converts a Weak channel into a Moderate one — the single highest-value action available on this application, and the variable the cohort has never yet tested.
- **If a letter is later wanted,** verified hooks are available: swissTAMP's reactive→proactive→predictive framing; the MUD condition-monitoring remit; the new-generation diagnostic-vehicle tender; and the honest thesis link (vibration-based condition monitoring — framed as a methods prototype, never as a validated system).
## Resume Plan — CONFIRMED BY USER 2026-08-21
### User decisions
1. **Role type + bundle: CONFIRMED.** Primary `bundle_data_engineer.md` (Staff/Senior Data Engineer, Tier 1). Secondary `bundle_analytics_engineer.md`, used only to reframe BS-5 (Spotfire) as the Power BI bridge.
2. **Language: ENGLISH**, with an explicit **Languages line stating German as mother tongue** — user's direct instruction. Because the document is English and carries no photo, the **International Tech** profile (config default, 2 pages) is the coherent choice; the one Swiss/DACH element retained is that Languages line, which the user asked for and which answers the obvious recruiter question on a German-language posting.
- Languages line: **German (native) · English (fluent)**. Norwegian/Russian (basic) optional — recommend omitting as non-professional.
- **Never list French or Italian** (config.md KB correction — Swisscom Zeugnis boilerplate, not accurate).
3. **Framing: delegated to Claude.** Chosen: lead on *operate-and-own*, carry the industrial sensor/asset-data thread from Bosch to now, drop GenAI/LLM material entirely.
### Format
- International Tech, **2 pages**, English, no photo, no demographic data
- Summary: 2–3 lines — operator identity plus the sensor/asset-data thread
- Skills: 4–6 compact lines, canonical `allowed` / `allowed-with-context` only
- Bosch title for this JD: **Senior Engineer, Data Analysis (Data Engineering)** — official title preserved with a transparent parenthetical per `resume_reference.md` §8
- Swisscom promotion shown: Senior → Staff (Engineer IV), April 2025
### Bullet plan — 11 bullets
**Swisscom — Staff Data, Analytics & AI Engineer (Oct 2023 – Present)** — budget 4–5, selected **5**
| | ID | Achievement | JD match | Why |
|---|---|---|---|---|
| * | SW-2 | Component Owner, Fulfillment ETL (Oracle → Kafka → Teradata), on-call SLA, governance | **Direct** R2, P1, P5 | **Lead bullet.** "Betrieb" and "Verantwortung für Qualität und Nachhaltigkeit" are the JD's own words; this is the closest evidence in the KB. |
| * | SW-1 | Migrated his domains' Teradata/Oracle ETL to AWS (S3, Glue, Athena/Iceberg, Redshift, Airflow, CloudFormation) | **Direct** R3, R4, P1 | Cloud data-platform build and operation. Scope-corrected: **his domains'**, contributor to the wider programme. |
| * | SW-7 | Governed data products and active metadata (Atlassian Compass) within the company-wide Data Mesh | **Direct** P3, R5 | The only direct answer to "Datenprodukte und semantische Modelle". Never claim ownership of the mesh. |
| * | SW-3 | Python data applications on Kubernetes with GitLab CI/CD | **Direct** R6 | Covers the required Python / tests / CI-CD / deployment line outright. |
| o | SW-4 | B2B data products and dashboards; process automation; RCA under 2nd/3rd-level support | **Direct** P5, P2 | Covers "Anforderungen mit dem Business aufnehmen". **First to cut** if space is tight. |
| x | SW-6 | PySpark | — | Fold into Skills, not a bullet. |
| x | SW-5 | Security Champion | — | JD has no security requirement. Default OMIT stands. |
| x | SW-8 | Domain-grounded LLM agents | — | Off-thesis here, long forbidden list. Dropped per framing. |
**Bosch — Senior Engineer, Data Analysis (Feb 2020 – Dec 2022)** — budget 3–4, selected **4**
| | ID | Achievement | JD match | Why |
|---|---|---|---|---|
| * | BS-2 | Data services in Python/Java/C# over OracleDB and Hadoop/Impala — fab sensor and process data: defect-management records, wafer inspection images, PCM electrical parameters | **Adjacent (strong)** P2 | **The differentiator.** Genuine industrial sensor/asset data for a role whose purpose includes *vorausschauende Wartung*. Also carries the Java thread. |
| * | BS-3 | Application Owner, analytics suite — SLOs, training, documentation, vendor management | **Direct** R1, R2 | Ownership plus the training/knowledge-transfer evidence R1 requires. |
| * | BS-5 | Co-owned TIBCO Spotfire platform, built C# extensions, co-presented TAF 2022 | **Bridge** R5 | **Carries the Power BI gap** — the weakest required line. A co-owned BI platform plus a public speaking credential is the honest answer. |
| * | BS-4 | ELK + Kafka anomaly-detection **PoC**, Grafana/Prometheus/Loki monitoring | **Bridge** P2 | Closest evidence to condition monitoring on streaming data. **PoC label must be retained.** |
| o | BS-1 | Containerized ML inference (Docker/K8s/Ansible) into 24/7 fab, automated defect classification | Adjacent | Strong bullet, but the JD wants pipelines that *enable* predictive maintenance, not ML delivery. Swap in for BS-4 if the production-ML angle is preferred. |
**Vizrt — Test Automation / DevOps Engineer (Jul 2017 – May 2018)** — selected **1**
| | ID | Achievement | JD match | Why |
|---|---|---|---|---|
| * | VZ-2 | Python A/V integration and unit test suite wired as CI/CD quality gates | **Direct** R6 | Reinforces the required tests/CI-CD line with a second, independent example. |
| x | VZ-1 | Distributed video transcoding backend | — | C++ — off-stack, and C++ is de-emphasized by standing preference. |
**Generali — IT Consultant (May 2015 – Jun 2017)** — selected **1**
| | ID | Achievement | JD match | Why |
|---|---|---|---|---|
| * | GN-1 | Introduced BDD; technical ownership of the automation suite; Jenkins administration; trained colleagues and presented to the Java Community | **Direct** R1 | R1 is a *required* line about developing other engineers. This is the clearest "introduced a practice and taught it" evidence in the KB. |
| o | GN-3 | Java/J2EE workflow features; WebServices to XLDeploy; **Apache Camel + Spring Boot PoC** | Bridge P4 | Only if the user wants to signal willingness for the RAMSI web-app work. See open question below. |
**Fraunhofer — 0 bullets.** FC-1 is C#/.NET plus Jenkins, off-stack here.
> **FC-4 deliberately excluded.** "Predictive Maintenance Research Grant Contribution" is a *grant proposal* contribution ("Mitarbeit"), not delivered work, and it is **absent from `claims.json`**. Despite being the most tempting keyword match in the entire KB for this JD, it must not be used as a predictive-maintenance credential — CLAUDE.md Rule 5 and the anti-fabrication rules both bar it.
**Total: 11 bullets** (Swisscom 5, Bosch 4, Vizrt 1, Generali 1) — bottom of the 11–14 guide, which is right here; padding to 14 would mean weaker evidence.
### Skills lines (draft, canonical only)
- **Languages & Data:** Python · SQL · PySpark · Java (Bosch/Generali) · C# (Bosch/Fraunhofer)
- **Pipelines & Streaming:** Apache Kafka · Apache Airflow · ETL/ELT · Oracle · Teradata · Hadoop/Impala
- **Cloud & Platform:** AWS (S3 · Glue · Athena/Iceberg · Redshift · Lambda · Step Functions) · CloudFormation/IaC · Kubernetes · Docker · GitLab CI/CD
- **Data Products & Governance:** Data Mesh data products · active metadata and lineage (Atlassian Compass) · data governance · data modelling
- **Analytics & Monitoring:** TIBCO Spotfire (co-owned, C# extensions) · Grafana · Prometheus · ELK (PoC)
- **Languages:** German (native) · English (fluent)
> **Explicitly absent, by design:** Snowflake, dbt, Argo Workflows, Helm, Power BI, Spring Boot, Angular, Terraform, Azure, GCP. Never substitute a JD word for an evidenced tool.
### Open question for the user (does not block Phase 2)
**Spring Boot.** `experience_generali.md` GN-3 records a contribution to an *Apache Camel + Spring Boot Dispatcher PoC* (2015–17). Canonical GN-3 records only "contributed to integration proofs of concept", and **Spring Boot is absent from the canonical skills list**, so it stays out of the document either way — a 9-to-11-year-old PoC is not a resume claim. But if the memory is accurate it is a legitimate *interview* answer to the RAMSI question. Worth confirming.
## Reuse Safety & Phase 2 Gates
- **`historical_outputs.json` checked 2026-08-21.** No prior package is reusable here and none is a match in kind: every existing output folder is an **English International Tech** resume, and six are on the `unsafe_do_not_reuse` list. **This would be the first Swiss/DACH German-language package in the repo** — build fresh from canonical claims, per policy.
- **Phase 2 forbidden-token gate (must not be skipped):** after generating the `.tex`, grep it for `Snowflake`, `dbt`, `Power BI`, `Spring Boot`, `Angular`, `Argo`, `Helm`, `LangChain`, `Terraform`, `Azure`, `GCP`. All must be **absent**. They appear throughout *this* session file because it analyses them as gaps — that is correct here and wrong in the document.
- Standard gates still apply: `validate_resume_system.py --document`, compile with MiKTeX pdflatex at full path, inspect the rendered PDF.
## Open Risks (surface to the user before Phase 1)
1. **Comp (C4) — the deciding factor, and it is unresolved.** Anforderungsniveau K is a collectively-bargained band with a fixed ceiling; the top of the band is not negotiable the way a private offer is. The current K figures could not be retrieved from public sources on 2026-08-21. Against the standing bar (Swisscom ~140k, moves only at 180k+ all-in) this is the most likely reason not to proceed, and it is a question for the People Leader, not for further desk research.
2. **Level (lateral or below).** Senior IC seat vs current Staff + Component Owner scope. The Bern/Thun WLB-exception tier deliberately accepts below-bar *comp*; it has never accepted below-*level* scope, and BKW was declined on exactly this. Worth asking whether the role carries design/architecture ownership in practice.
3. **P4 Java/Spring Boot/Angular.** A full responsibility line is web-application development on a stack that is historical (Java) or absent (Spring Boot, Angular). Not a gate, but it will come up, and it is work he may not want.
4. **Five named products in the required stack are non-canonical.** Defensible because of "z. B.", but the resume will visibly not say Snowflake, dbt or Power BI. An ATS keyword screen configured on those literals would drop him. Nothing can honestly be done about that.
## Output Files
- Resume: `e2e_sbb_data_engineer_asset_mgmt_resume.tex` → `.pdf` (**2 pages**, generated 2026-08-21)
- Cover letter: **intentionally omitted** — SBB explicitly waives it
- Critique: `critique_sbb_data_engineer_asset_mgmt.md` (2026-08-21)
## Critique Summary
- Evidence Fit: **79/100 — Core (lower end)**, unchanged; a document cannot move candidate-role fit
- Hard gate: **PASS**
- Document Quality: **92/100** after Edit 1 (was 90 at critique time) — truth/provenance 24/25, well clear of the 8/10 automatic-failure line
- Channel Strength: **Weak**, but uniquely upgradeable — the first posting in the log with a named hiring contact and a direct number
- **Tier 1 fix (one):** the JD names *Abfrageleistungsoptimierung* in a required line and the document is silent on query performance, despite canonical support in BS-2's 3L variant. The only substantive gap that costs nothing in honesty.
- Tier 2: "agile" **RESOLVED** (Edit 1, SW-3). APIs **not applied** — no canonical evidence that source onboarding used APIs; needs user confirmation before it can be added.
- **No Tier 1 truth findings.** Claim audit clean across all 13 bullets; SW-1 and SW-7 both avoid every forbidden phrasing.
- Reader sequence: ATS pass (fails only a screen keyed on the three unnameable products) · Recruiter **Forward** · Hiring manager **Interview, probable** · Technical reviewer **Credible**.
- **Verdict: the document is ready; the decision is the constraint.** Do not submit before the Anforderungsniveau K answer.
- Framework note: the critique skill's single 8-dimension score (incl. Publications) conflicts with `critique_framework.md` and CLAUDE.md, which mandate three separate scores. Followed CLAUDE.md.
### Phase 2 record
- **13 bullets** shipped, not the 11 planned. Two deliberate changes during generation:
1. **Fraunhofer kept 1 bullet (FC-1 + FC-3 combined), not 0.** Dropping the role would have opened an unexplained **21-month gap** (Vizrt ends May 2018, Bosch starts Feb 2020). Chronology integrity outranks the bullet budget.
2. **BS-1 added** as a fifth Bosch bullet (marked `o`/available in the plan). Not padding: it is HIGH priority in `bundle_data_engineer.md`, and containerized delivery into a *continuously operating* environment reads directly onto an employer that runs a 24/7 network. It also improved page-2 fill.
- **Page-break decision.** Tested the natural flow with `\newpage` removed: page 1 filled completely, but two Bosch bullets then landed on page 2 with no visible employer header — an information-hierarchy defect under `resume_reference.md` §3/§9. Kept the guarded break so every page's content sits under its own header. Remaining bottom whitespace is acceptable per §4 ("white space is acceptable when hierarchy and readability are strong") and the 2026-07-27 KB correction removing page-fill quotas.
- **Skills line relabelled** — the first line read "Languages" (programming) and the last "Languages spoken", which scanned as a duplicate. Now **Programming** and **Languages**.
- **Gates:** canonical validator **PASS** (0 warnings) · forbidden-token scan **PASS** (none of the seven non-canonical JD tools present) · scope-discipline scan **PASS** (no ownership verb paired with an org-scale object) · compile **PASS** (exactly 2 pages) · both pages visually inspected — no clipping, overlap, orphaned heading or bad break; umlauts render correctly.
- **Note for future runs:** the validator scans **raw file text including comments**. A header comment listing the banned tool names failed the gate on the first pass. Describe them; never name them.
## Edit History
### Edit 1 (2026-08-21): critique Tier 1 + one Tier 2
- **Source:** `critique_sbb_data_engineer_asset_mgmt.md` items 1 and 2
- **Changes (all MODIFY — no bullet added, removed or swapped):**
1. BS-2 bullet: "giving analysis teams access to…" → "**tuning query performance for** analysis teams working with…". Canonical support: BS-2 3L variant, "optimized query performance and data availability".
2. Pipelines skills line: added "query performance tuning".
3. SW-3 bullet: added "**in an agile DevOps team**". Canonical support: SW-3 2L variant.
- **Not applied:** the API term (critique Tier 2 item 3). No canonical evidence that source onboarding used APIs — adding it would be vocabulary substitution. Needs user confirmation.
| Metric | Before | After | Delta |
|---|---:|---:|---|
| Page count | 2 | 2 | 0 |
| Bullets | 13 | 13 | 0 |
| Validator | PASS | PASS | 0 |
| Forbidden tokens | 0 | 0 | 0 |
| "query" occurrences | 0 | **2** | +2 |
| "agile" occurrences | 0 | **1** | +1 |
| Longest bullet (words) | 31 | 34 | +3 (still under the 35 flag) |
- **Verification:** validator PASS (0 warnings) · forbidden-token scan PASS · compile PASS at 2 pages · both pages re-rendered and inspected — no clipping, overlap, orphan or bad break.
- **Effect on scores:** Document Quality relevance/terminology moves 11→13, so **Document Quality 90 → 92/100**. Evidence Fit unchanged at 79 (edits do not change candidate-role fit).
## Status
- Fit gate: **DONE — PASS, Core, Evidence Fit 79/100**
- Phase 0: **DONE** (2026-08-21)
- Phase 1: **DONE** (2026-08-21) — user confirmed bundle, chose English with German-native stated, delegated framing
- Resume: **DONE** (2026-08-21, Edit 1 applied) — 13 bullets, 6 skills lines, 2 pages, all gates PASS
- Cover letter: NO (SBB waives it)
- Critique: **DONE** (2026-08-21) — Document Quality 92/100 after Edit 1; **no outstanding Tier 1 fixes**
- Application/outcome: **not submitted — held pending the comp answer**
- Next: call Andri Wienandts (+41 79 364 62 53) on the Anforderungsniveau K band and the seat’s design scope → then either apply the Tier 1 query-performance fix and submit, or close as NO-GO on comp
resume_builder/bundles/bundle_data_engineer.md
+3 -1
View File
@@ -102,7 +102,9 @@ Python, Kafka, AWS (S3 · Glue · Athena · Redshift · Airflow · CloudFormatio
**Institution type:** Industry — tech company, scale-up, or enterprise platform team
**Opening hook pattern:**
> "As a Staff Data Engineer at Swisscom — Switzerland's largest telco — I currently own the business-critical Fulfillment ETL pipelines that feed our data warehouse, while simultaneously leading the migration of our legacy stack to a cloud-native AWS architecture. [Tie to their specific need / JD signal]."
> "As a Staff Data Engineer at Swisscom — Switzerland's largest telco — I currently own the business-critical Fulfillment ETL pipelines that feed our data warehouse, and I migrated my own domains' pipelines onto the company's AWS platform as part of its wider cloud migration programme. [Tie to their specific need / JD signal]."
> **Scope warning (corrected 2026-08-21).** This hook previously read "*leading the migration of our legacy stack*", which contradicts both `claims.json` SW-1 (`forbidden`: "led the migration of the legacy warehouse", "sole technical lead") and point 2 of the narrative thread directly below. Never pair a full-ownership verb with a company-scale object. See CLAUDE.md → Scope Discipline.
**Key narrative thread:**
1. **Ownership at scale** — Component Owner at Swisscom, Application Owner at Bosch: not just building pipelines, but running them in production with SLA accountability
resume_builder/bundles/bundle_data_platform.md
+9 -2
View File
@@ -106,11 +106,18 @@ Kubernetes, Docker, AWS (S3 · Glue · Athena · Redshift · CloudFormation), Ka
**Institution type:** Cloud-first tech company, scale-up with AWS-heavy stack, enterprise platform team, or data infrastructure consultancy
**Opening hook pattern:**
> "Across my career at Swisscom and Bosch, I've owned data infrastructure at two ends of the spectrum: migrating Swisscom's legacy ETL stack to a cloud-native AWS platform (CloudFormation, Glue, Athena with Iceberg, Airflow) while operating Kubernetes-deployed Python applications with GitLab CI/CD — and containerizing ML inference into a 24/7 semiconductor production line at Bosch using Docker, Kubernetes, and Ansible. In both cases, the infrastructure had to be production-grade with no tolerance for downtime. [Tie to their platform challenge]."
> "Across my career at Swisscom and Bosch, I've owned data infrastructure at two ends of the spectrum: migrating my own domains' pipelines onto Swisscom's AWS platform (CloudFormation, Glue, Athena with Iceberg, Airflow) as part of its wider cloud programme, while operating Kubernetes-deployed Python applications with GitLab CI/CD — and containerizing ML inference into a 24/7 semiconductor production line at Bosch using Docker, Kubernetes, and Ansible. In both cases, the infrastructure had to be production-grade with no tolerance for downtime. [Tie to their platform challenge]."
> **Scope warning (corrected 2026-08-26).** This hook previously read "*migrating Swisscom's legacy
> ETL stack to a cloud-native AWS platform*" — a full-ownership framing on a company-scale object,
> contradicting `claims.json` SW-1 (`forbidden`: "migrated the company warehouse", "led the migration
> of the legacy warehouse"). Same defect fixed in `bundle_data_engineer.md` on 2026-08-21; this
> bundle was missed by that sweep. Never pair a full-ownership verb with a company-scale object.
> See CLAUDE.md → Scope Discipline.
**Key narrative thread:**
1. **Production Kubernetes** — SW-3 + BS-1: K8s at two employers, in different contexts (data apps at Swisscom, ML inference at Bosch). Cross-employer K8s ownership is a strong signal.
2. **Full AWS platform stack** — SW-1: Not just using one AWS service — migrating an entire ETL infrastructure to AWS with CloudFormation/IaC shows platform-level thinking.
2. **AWS platform breadth** — SW-1: not just one AWS service — migrating his own domains' pipelines with CloudFormation/IaC, inside the wider company programme, shows platform-level thinking. **Never write "an entire ETL infrastructure" or any company-scale object here** (corrected 2026-08-26).
3. **Observability initiative** — BS-4: Self-initiated ELK + Prometheus + Grafana PoC shows platform engineer mindset (monitoring is not optional).
4. **Operational accountability** — use Component/Application Owner evidence. Security Champion is optional and never a substitute for production ownership.
resume_builder/bundles/bundle_ml_ai_engineer.md
+1 -1
View File
@@ -59,7 +59,7 @@
|----|----------------|--------------------|--------------------|
| BS-1 | ML inference containerization | **LEAD bullet** — "Designed and deployed ML inference pipeline (Docker, K8s, Ansible) into 24/7 semiconductor fab; automated image-based defect classification" | Production ML in constrained 24/7 environment |
| SW-3 | K8s + GitLab CI/CD | "Deployed and operated ML-ready Python applications on Kubernetes with GitLab CI/CD automation — production-grade containerized delivery" | K8s ownership = MLOps infrastructure |
| SW-1 | AWS migration | "Built cloud-native data infrastructure on AWS (S3, Glue, Athena/Iceberg, Redshift, Airflow) — the scalable data layer that ML models depend on" | AWS data lake for ML workloads |
| SW-1 | AWS migration | "Migrated his domains' pipelines onto Swisscom's AWS platform (S3, Glue, Athena/Iceberg, Redshift, Airflow) — the data layer ML workloads depend on" — **corrected 2026-08-26**, previously read "Built cloud-native data infrastructure… the scalable data layer", pairing a full-ownership verb with an org-scale object (SW-1 `forbidden`) and asserting unverified scale | AWS data lake for ML workloads |
| SW-2 | Component Owner | "Owned business-critical ETL pipelines (Oracle/Kafka → Teradata) — reliable data supply for downstream ML and analytics" | Data reliability for ML input |
| FC-2 | ARTUS NLP | "Contributed ML and speech recognition components to ARTUS — Fraunhofer research project targeting automatic sea rescue transcription" | Applied NLP in safety-critical domain |
| BS-4 | ELK PoC | "Delivered anomaly detection PoC: ELK + Kafka pipeline with Grafana/Prometheus monitoring — ML-adjacent signal processing" | Anomaly detection / observability |
resume_builder/canonical/claims.json
+497 -73
View File
@@ -1,9 +1,15 @@
{
"schema_version": 1,
"last_verified": "2026-07-27",
"last_verified": "2026-08-27",
"authority": {
"description": "Machine-readable source of truth for all future application documents.",
"precedence": ["canonical claims", "config corrections", "verified references", "experience files", "bundles"],
"precedence": [
"canonical claims",
"config corrections",
"verified references",
"experience files",
"bundles"
],
"raw_extractions_are_immutable": true,
"historical_outputs_are_never_sources": true
},
@@ -19,8 +25,18 @@
"swiss_work_authorization": "Authorized to work in Switzerland; no visa or employer sponsorship required",
"work_authorization_source": "User-confirmed 2026-07-27",
"resume_usage": "Work authorization is optional in the header; mention the B permit or no-sponsorship status when it resolves recruiter uncertainty or the application asks for it",
"languages": {"German": "native", "English": "fluent", "Norwegian": "basic", "Russian": "basic"},
"forbidden_personal_data": ["date of birth", "marital status", "gender", "children"]
"languages": {
"German": "native",
"English": "fluent",
"Norwegian": "basic",
"Russian": "basic"
},
"forbidden_personal_data": [
"date of birth",
"marital status",
"gender",
"children"
]
},
"education": [
{
@@ -32,8 +48,10 @@
"end": "2013-10",
"thesis_institution": "Tongji University, Shanghai",
"thesis_title": "Development of a Web-Based Remote Fault Diagnosis System",
"final_grade": "1.6 (German scale, 1.0 best / 4.0 lowest pass) -- user-confirmed 2026-08-26",
"final_grade_us_4pt": "3.4 -- modified Bavarian formula, US = 1 + 3*(4.0-G)/(4.0-1.0), which simplifies to 5.0 - G. Use ONLY when a form demands a 4.0-scale GPA; always state the German scale alongside it. Never present as a natively-awarded GPA.",
"thesis_grade": "1.0",
"thesis_relative_grade": "ECTS B (top 35%); official certificate on file",
"thesis_relative_grade": "ECTS grade B; official certificate on file. B is formally the next 25% after the top 10%, i.e. the 10-35% band -- so the correct and safe phrasing is 'top 35%' (user-confirmed 2026-08-26). Do NOT write 'top 30%' or 'top 25%': B is a band, not a point, and narrowing it overstates the standing. The original 'top 35%' in this record was correct; a briefly-considered 30% was the error.",
"thesis_methods": "Verified against the thesis PDF 2026-08-02. Domain: vibration-based condition monitoring of CNC machine tools (piezoelectric accelerometers on spindle, tool rest, lathe body). Features: speed, load and the dimensionless waveform, peak, pulse, margin and kurtosis indices. Surveyed CBR, PSO, RBR and ANN; SELECTED a hybrid of rule-based reasoning (interpretable thresholds) plus a 7-10-3 feed-forward ANN (faultstates green/yellow/red) to handle noisy data that static rules cannot classify. Extensible plug-in architecture for swappable diagnosis methods and data collectors. Java/GWT client-server on MySQL. Evaluation measured throughput and latency, NOT model accuracy: the pipeline sustained ~500 samples/s against tri-axial vibration sensors delivering 72.9 kSPS, so the thesis concluded data reduction, filtering or batch scheduling is required for real-time use.",
"thesis_limits": "PSO was SURVEYED ONLY and NOT used in the implementation -- never claim it as an applied method. No real operational data: training data came from a cited thesis, test data was modified plus a random generator. No model-accuracy figures exist. Back-propagation retraining was listed as future work, not implemented. Frame as a methods prototype/framework, never as a validated production predictive-maintenance system.",
"transcripts_language": "English-language originals available for B.Eng. and M.Eng."
@@ -43,7 +61,10 @@
"degree": "B.Eng. Information and Telecommunication Technologies",
"institution": "Universitat der Bundeswehr Munchen",
"start": "2009-10",
"end": "2012-10"
"end": "2012-10",
"final_grade": "2.4 (German scale, 1.0 best / 4.0 lowest pass) -- user-confirmed 2026-08-26",
"final_grade_us_4pt": "2.6 -- modified Bavarian formula (5.0 - G). Supply ONLY when a form requires a bachelor GPA; leave blank where optional. Always state the German scale alongside it, because the conversion makes German degrees read weaker to a US audience than they are: 2.4 is a solid 'gut'. Never present as a natively-awarded GPA.",
"grade_output_rule": "Neither degree grade goes on the resume or in a cover letter by default. Industry resumes omit grades; the M.Eng. thesis grade 1.0 is the only figure worth surfacing, and only where academic performance is genuinely valued. These fields exist for application-form GPA questions and credential checks."
}
],
"employment": [
@@ -55,8 +76,16 @@
"start": "2023-10",
"end": "present",
"title_history": [
{"title": "Senior Data, Analytics & AI Engineer", "start": "2023-10", "end": "2025-04"},
{"title": "Staff Data, Analytics & AI Engineer", "start": "2025-04", "end": "present"}
{
"title": "Senior Data, Analytics & AI Engineer",
"start": "2023-10",
"end": "2025-04"
},
{
"title": "Staff Data, Analytics & AI Engineer",
"start": "2025-04",
"end": "present"
}
]
},
{
@@ -64,9 +93,15 @@
"employer": "Robert Bosch Semiconductor Manufacturing Dresden GmbH",
"location": "Dresden, Germany",
"official_title": "Engineer / Senior Engineer, Data Analysis",
"display_title": "Senior Engineer, Data Analysis (Data & ML Engineering)",
"display_title": "Senior Data Engineer, Data Analysis",
"display_title_note": "experience_bosch.md sanctions title flexibility for this role: '(Senior) Data Engineer' or '(Senior) Data Analysis Engineer' depending on the JD. 'Senior Data Engineer, Data Analysis' is the user's own preferred wording, confirmed 2026-08-27. official_title stays the formal combined string on the Zeugnis.",
"start": "2020-02",
"end": "2022-12"
"end": "2022-12",
"title_history": [
{"title": "Data Engineer, Data Analysis", "start": "2020-02", "end": "2021-01"},
{"title": "Senior Data Engineer, Data Analysis", "start": "2021-01", "end": "2022-12"}
],
"title_history_source": "LinkedIn-confirmed, recorded in experience_bosch.md; user-confirmed 2026-08-27 while reviewing the RUAG C5I resume. Added because claims.json previously carried only the combined official_title with no promotion date, which made a Swisscom-style 'Beforderung <date>' line unwritable from the canonical file alone."
},
{
"id": "FRAUNHOFER",
@@ -101,178 +136,567 @@
"display_title": "Software Engineer",
"start": "2014-11",
"end": "2015-05"
},
{
"id": "BUNDESWEHR",
"employer": "Bundeswehr (German Armed Forces)",
"location": "Germany",
"display_title": "Officer Candidate / Officer (Second Lieutenant at separation)",
"start": "2008-07",
"end": "2014-11",
"source": "User confirmation 2026-07-10; thiessen_linkedin_profile.md",
"last_verified": "2026-08-27"
}
],
"claims": [
{
"id": "SW-1",
"scope": "Primary engineer for pipelines in owned Fulfillment and Product Analysis domains; contributor to the wider company migration.",
"allowed_verbs": ["migrated", "implemented", "contributed"],
"forbidden": ["led the migration of the legacy warehouse", "sole technical lead", "migrated the company warehouse", "owned the full migration"],
"allowed_verbs": [
"migrated",
"implemented",
"contributed"
],
"forbidden": [
"led the migration of the legacy warehouse",
"sole technical lead",
"migrated the company warehouse",
"owned the full migration"
],
"metrics": "unverified"
},
{
"id": "SW-2",
"scope": "Component Owner for business-critical Fulfillment ETL pipelines, including operation, data quality, governance, incidents and on-call obligations.",
"allowed_verbs": ["own", "operate", "maintain", "serve"],
"allowed_verbs": [
"own",
"operate",
"maintain",
"serve"
],
"metrics": "unverified"
},
{
"id": "SW-3",
"scope": "Build and operate Python data applications on Kubernetes with GitLab CI/CD within the team environment.",
"allowed_verbs": ["build", "deploy", "operate"],
"allowed_verbs": [
"build",
"deploy",
"operate"
],
"metrics": "unverified"
},
{
"id": "SW-4",
"scope": "Deliver data products, dashboards and analyses with internal B2B stakeholders and product owners; not external consulting or strategic-account delivery.",
"allowed_verbs": ["deliver", "translate", "partner", "support"],
"forbidden": ["customer-embedded delivery", "strategic customer delivery", "C-suite advisor"],
"allowed_verbs": [
"deliver",
"translate",
"partner",
"support"
],
"forbidden": [
"customer-embedded delivery",
"strategic customer delivery",
"C-suite advisor"
],
"metrics": "unverified"
},
{
"id": "SW-5",
"scope": "Mandatory rotating team Security Champion role for 2025/2026 only; not an award, certification or multi-year distinction.",
"allowed_verbs": ["serve"],
"forbidden": ["3 consecutive years", "2023-2026", "Security Champion x3", "owning DevSecOps compliance"],
"allowed_verbs": [
"serve"
],
"forbidden": [
"3 consecutive years",
"2023-2026",
"Security Champion x3",
"owning DevSecOps compliance"
],
"metrics": "100 hours of training and an assessment are source-backed; use only when relevant"
},
{
"id": "SW-6",
"scope": "Hands-on PySpark use at Swisscom; scale and performance metrics are not verified.",
"allowed_verbs": ["use", "develop", "process"],
"allowed_verbs": [
"use",
"develop",
"process"
],
"metrics": "unverified"
},
{
"id": "SW-7",
"scope": "Build governed data products and onboard sources within Swisscom's company-wide Data Mesh; never claim ownership or construction of the shared mesh/platform. Atlassian Compass is the metadata/catalogue platform used for data-product metadata and lineage (user-confirmed 2026-07-29); Dennis uses it as a practitioner and does not administer or own the tooling.",
"allowed_verbs": ["build", "model", "onboard", "contribute", "document", "maintain"],
"forbidden": ["built a Data Mesh", "built the Data Mesh", "built AWS Data Mesh", "own the AWS data platform", "own the data platform", "own Atlassian Compass", "administered Compass", "rolled out Compass"],
"allowed_verbs": [
"build",
"model",
"onboard",
"contribute",
"document",
"maintain"
],
"forbidden": [
"built a Data Mesh",
"built the Data Mesh",
"built AWS Data Mesh",
"own the AWS data platform",
"own the data platform",
"own Atlassian Compass",
"administered Compass",
"rolled out Compass"
],
"metrics": "unverified"
},
{
"id": "SW-8",
"scope": "Configured domain-grounded LLM assistants in a Swisscom-owned web interface by selecting available models and supplying curated knowledge. Separate exposure includes LiteLLM API use, custom GPTs, Copilot and Kiro.",
"allowed_verbs": ["configured", "used", "integrated"],
"forbidden": ["built production LLM systems", "deployed LLM systems", "built LangChain", "agent orchestration", "formal LLM evaluation", "fine-tuned models"],
"allowed_verbs": [
"configured",
"used",
"integrated"
],
"forbidden": [
"built production LLM systems",
"deployed LLM systems",
"built LangChain",
"agent orchestration",
"formal LLM evaluation",
"fine-tuned models"
],
"metrics": "unverified"
},
{
"id": "BS-1",
"scope": "Designed and executed integration of containerized ML inference into a 24/7 semiconductor production environment; model-development ownership is not established.",
"allowed_verbs": ["integrated", "containerized", "deployed", "orchestrated"],
"forbidden": ["trained the image classification model", "owned the full ML lifecycle"],
"scope": "Designed and executed integration of containerized ML inference into a 24/7 semiconductor production environment; model-development ownership is not established. Toolchain includes Docker, Kubernetes and Ansible - Ansible use at Bosch was intensive (user-confirmed 2026-08-27) and extended to configuration management and infrastructure automation beyond this single integration, making it standalone Infrastructure-as-Code evidence.",
"allowed_verbs": [
"integrated",
"containerized",
"deployed",
"orchestrated"
],
"forbidden": [
"trained the image classification model",
"owned the full ML lifecycle"
],
"metrics": "qualitative reduction in manual classification is source-backed; exact amount unverified"
},
{
"id": "BS-2",
"scope": "Developed data services in Python, Java and C# over Oracle and Hadoop/Impala for internal analysis teams. Data types worked on (user-confirmed 2026-08-02): semiconductor fab sensor and process data -- defect management records, wafer inspection images, and electrical parameters from Process Control Monitoring (PCM). Relevant as genuine industrial sensor/asset-data experience.",
"allowed_verbs": ["developed", "built"],
"allowed_verbs": [
"developed",
"built"
],
"metrics": "unverified"
},
{
"id": "BS-3",
"scope": "Confirmed Application Owner responsibilities for analytics applications and upstream pipelines, including SLOs, vendors, training and documentation.",
"allowed_verbs": ["served", "owned", "managed", "defined"],
"allowed_verbs": [
"served",
"owned",
"managed",
"defined"
],
"metrics": "unverified"
},
{
"id": "BS-4",
"scope": "Built an ELK/Kafka anomaly-detection proof of concept and monitoring; not a company-wide observability platform.",
"allowed_verbs": ["built", "implemented", "validated"],
"forbidden": ["built the observability platform", "enterprise observability platform"],
"allowed_verbs": [
"built",
"implemented",
"validated"
],
"forbidden": [
"built the observability platform",
"enterprise observability platform"
],
"metrics": "unverified"
},
{
"id": "BS-5",
"scope": "Co-owned the TIBCO Spotfire environment, built C# extensions and co-presented at TIBCO Analytics Forum 2022.",
"allowed_verbs": ["co-owned", "built", "co-presented"],
"allowed_verbs": [
"co-owned",
"built",
"co-presented"
],
"metrics": "unverified"
},
{
"id": "BS-6",
"scope": "Administered and automated the Linux server estate underpinning fab analytics and ML workloads at Bosch (2020-2022): ML platform hosts, Docker hosts, backend services and the Ansible control path. Developed Ansible extensions, including a credential plugin that retrieved secrets from a password keystore and cached them locally to cut lookup volume and improve scalability and performance. Playbooks were version-controlled in Git and executed push-style from Jenkins/GitLab pipelines. User-confirmed 2026-08-27.",
"allowed_verbs": [
"administered",
"automated",
"developed",
"extended",
"operated",
"built"
],
"forbidden": [
"owned the fab's infrastructure",
"ran the datacenter",
"SRE role or title",
"GitOps",
"any server-count, uptime, SLA or scale figure - none is verified"
],
"metrics": "none verified; the credential-caching plugin reduced keystore lookups, amount unquantified"
},
{
"id": "FC-1",
"scope": "Set up Jenkins CI/CD and contributed to SCEDAS development and maintenance.",
"allowed_verbs": ["set up", "developed", "maintained"],
"allowed_verbs": [
"set up",
"developed",
"maintained"
],
"metrics": "unverified"
},
{
"id": "FC-2",
"scope": "Contributed ML/NLP components to ARTUS; no publication or model-training ownership.",
"allowed_verbs": ["contributed", "implemented", "supported"],
"forbidden": ["led ARTUS", "trained the speech model"],
"allowed_verbs": [
"contributed",
"implemented",
"supported"
],
"forbidden": [
"led ARTUS",
"trained the speech model"
],
"metrics": "unverified"
},
{
"id": "FC-3",
"scope": "Developed containerized microservices for the MISSION research platform.",
"allowed_verbs": ["developed", "built"],
"allowed_verbs": [
"developed",
"built"
],
"metrics": "unverified"
},
{
"id": "VZ-1",
"scope": "Contributed Python and C++ engineering to a distributed video-transcoding backend. Named broadcasters are product context, not direct delivery claims.",
"allowed_verbs": ["developed", "engineered", "contributed"],
"forbidden": ["for CNN, BBC and Al Jazeera", "delivered to CNN", "customer-embedded"],
"allowed_verbs": [
"developed",
"engineered",
"contributed"
],
"forbidden": [
"for CNN, BBC and Al Jazeera",
"delivered to CNN",
"customer-embedded"
],
"metrics": "unverified"
},
{
"id": "VZ-2",
"scope": "Developed automated audio/video integration tests and connected quality gates to CI/CD.",
"allowed_verbs": ["developed", "automated", "integrated"],
"allowed_verbs": [
"developed",
"automated",
"integrated"
],
"metrics": "unverified"
},
{
"id": "GN-1",
"scope": "Introduced BDD through a proof of concept and held technical responsibility for test automation, training and Jenkins jobs.",
"allowed_verbs": ["introduced", "owned", "trained", "administered"],
"allowed_verbs": [
"introduced",
"owned",
"trained",
"administered"
],
"metrics": "unverified"
},
{
"id": "GN-2",
"scope": "Developed UIPath RPA proofs of concept and acted as an internal contact.",
"allowed_verbs": ["developed", "served"],
"allowed_verbs": [
"developed",
"served"
],
"metrics": "unverified"
},
{
"id": "GN-3",
"scope": "Developed Java/J2EE workflow application features and contributed to integration proofs of concept.",
"allowed_verbs": ["developed", "contributed", "migrated"],
"allowed_verbs": [
"developed",
"contributed",
"migrated"
],
"metrics": "unverified"
},
{
"id": "BW-1",
"scope": "Completed officer candidate training and officer school during six years of service in the German Armed Forces (Bundeswehr), leaving service as Second Lieutenant. This establishes military and organisational context only; no technical military role, command scope, deployment, NATO service or clearance is established.",
"ownership": "Individual service and completed training.",
"allowed_verbs": [
"completed",
"served",
"left service"
],
"forbidden": [
"current or former security clearance",
"NATO service",
"combat role or deployment",
"technical AI, ICT or cyber work for the military",
"command responsibility or leadership outcomes not explicitly verified"
],
"metrics": "Six years of service and separation as Second Lieutenant are verified; no performance or command-scope metric is established.",
"source": "User confirmation 2026-07-10; thiessen_linkedin_profile.md",
"last_verified": "2026-08-27"
},
{
"id": "PP-1",
"scope": "PERSONAL PROJECT (user-supplied 2026-08-25), not professional work. A self-hosted, single-user investing/signal platform Dennis built for himself: ingests daily US-equity prices, fundamentals and sentiment (LLM-assisted); runs one long-only cross-sectional momentum book (top-quintile residual 12-1 momentum, ATR stop/trail, max 15 concurrent names) through scheduled scan and backtest pipelines; surfaces gated setups in a web dashboard plus Telegram alerts. Legitimate use: evidence of self-directed trading-domain fluency and of an end-to-end ingestion/backtest/alerting build outside work. Full-ownership verbs ARE correct here -- unlike his employer work, this is genuinely solo.",
"allowed_verbs": [
"built",
"developed",
"designed"
],
"forbidden": [
"any trading track record, PnL, return, Sharpe or backtest performance figure -- none is verified and none may ever be quoted",
"calling it distributed, scalable, production or multi-user -- it is single-user and self-hosted",
"implying professional quantitative, trading, finance-domain or investment-management experience",
"presenting it as employer work or omitting that it is a personal project"
],
"metrics": "unverified"
},
{
"id": "PP-2",
"scope": "Udacity 'AI for Trading' Nanodegree, completed self-study (user-supplied 2026-08-25). Covers quantitative trading fundamentals -- factor construction, alpha signals, backtesting. Pairs with PP-1 as the formal half of self-directed domain grounding.",
"allowed_verbs": [
"completed"
],
"forbidden": [
"listing it as an academic degree, professional certification or quant credential",
"presenting it as equivalent to quantitative research or trading experience",
"using it to qualify for quant, trader, researcher or portfolio-management roles"
],
"metrics": "unverified"
},
{
"id": "PP-3",
"scope": "PERSONAL PROJECT (user-supplied 2026-08-27), not employer work. A self-hosted Debian server Dennis has operated for ~10 years and administers and hardens himself, running nginx, a mail server, Nextcloud, a VPN server, Docker services and Bitwarden. This is the primary evidence for hands-on Linux administration, system hardening and security configuration, and it is current and continuous - unlike the Bosch Linux work, which ended in 2022. User states he genuinely enjoys this kind of work, including physical infrastructure.",
"allowed_verbs": [
"operate",
"administer",
"harden",
"run",
"maintain",
"self-host"
],
"forbidden": [
"calling it enterprise, production, multi-user or business-critical infrastructure",
"implying professional SRE, sysadmin or hosting employment",
"any uptime, availability, SLA, user-count or scale figure - none is verified",
"presenting it as employer work or omitting that it is self-hosted and personal",
"using it as evidence of datacenter or physical-infrastructure experience"
],
"metrics": "none; ~10 years of continuous operation is the only quantity, and it is self-reported"
}
],
"skills": [
{"name": "Python", "evidence": "production-current", "output": "allowed"},
{"name": "SQL", "evidence": "production-current", "output": "allowed"},
{"name": "PySpark", "evidence": "production-current", "output": "allowed"},
{"name": "Kafka", "evidence": "production-current", "output": "allowed"},
{"name": "Airflow", "evidence": "production-current", "output": "allowed"},
{"name": "AWS", "evidence": "production-current", "output": "allowed"},
{"name": "CloudFormation", "evidence": "production-current", "output": "allowed"},
{"name": "Atlassian Compass", "evidence": "production-current", "output": "allowed", "note": "Metadata/catalogue platform for data-product metadata and lineage at Swisscom (user-confirmed 2026-07-29). Practitioner use only — do not claim administration, rollout or ownership. Satisfies 'or similar metadata/catalogue platform' phrasing; is NOT a substitute claim for Purview, Collibra or Alation, which remain unevidenced."},
{"name": "Kubernetes", "evidence": "production-current-and-historical", "output": "allowed"},
{"name": "Docker", "evidence": "production-current-and-historical", "output": "allowed"},
{"name": "Java", "evidence": "production-historical", "output": "allowed-with-context"},
{"name": "C#", "evidence": "production-historical", "output": "allowed-with-context"},
{"name": "C++", "evidence": "production-historical-limited", "output": "allowed-with-context"},
{"name": "JavaScript", "evidence": "production-historical-limited", "output": "allowed-with-context"},
{"name": "LiteLLM", "evidence": "hands-on-current", "output": "allowed-with-context"},
{"name": "custom GPTs", "evidence": "hands-on-current", "output": "allowed-with-context"},
{"name": "Kiro", "evidence": "hands-on-current", "output": "allowed-with-context"},
{"name": "Copilot", "evidence": "hands-on-current", "output": "allowed-with-context"},
{"name": "TensorFlow/Keras", "evidence": "certification", "output": "certification-context-only"},
{"name": "PyTorch", "evidence": "coursework-or-personal-unverified", "output": "certification-context-only"},
{"name": "TypeScript", "evidence": "unverified", "output": "forbidden"},
{"name": "FastAPI", "evidence": "unverified", "output": "forbidden"},
{"name": "Flask", "evidence": "unverified", "output": "forbidden"},
{"name": "Django", "evidence": "unverified", "output": "forbidden"},
{"name": "LangChain", "evidence": "never-used", "output": "forbidden"},
{"name": "LangGraph", "evidence": "never-used", "output": "forbidden"},
{"name": "LlamaIndex", "evidence": "never-used", "output": "forbidden"},
{"name": "formal model evaluation", "evidence": "unverified", "output": "forbidden"},
{"name": "LLM fine-tuning", "evidence": "unverified", "output": "forbidden"},
{"name": "Azure", "evidence": "unverified", "output": "forbidden"},
{"name": "GCP", "evidence": "unverified", "output": "forbidden"},
{"name": "Terraform", "evidence": "unverified", "output": "forbidden"}
{
"name": "Python",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "SQL",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "PySpark",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "Kafka",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "Airflow",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "AWS",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "CloudFormation",
"evidence": "production-current",
"output": "allowed"
},
{
"name": "Atlassian Compass",
"evidence": "production-current",
"output": "allowed",
"note": "Metadata/catalogue platform for data-product metadata and lineage at Swisscom (user-confirmed 2026-07-29). Practitioner use only — do not claim administration, rollout or ownership. Satisfies 'or similar metadata/catalogue platform' phrasing; is NOT a substitute claim for Purview, Collibra or Alation, which remain unevidenced."
},
{
"name": "Kubernetes",
"evidence": "production-current-and-historical",
"output": "allowed"
},
{
"name": "Ansible",
"evidence": "production-historical",
"output": "allowed",
"note": "User-confirmed 2026-08-27: worked intensively with Ansible at Bosch (2020-2022) for configuration management and infrastructure automation, including the BS-1 ML-inference orchestration. Was absent from claims.json despite already appearing in experience_bosch.md BS-1 bullet variants. Counts as genuine Infrastructure-as-Code evidence alongside CloudFormation. Does NOT by itself license claiming Terraform (still unverified) or GitOps as a named practice (see gitops_pending)."
},
{
"name": "Linux administration",
"evidence": "production-historical-and-personal-current",
"output": "allowed",
"note": "User-confirmed 2026-08-27 and previously ABSENT from claims.json entirely. Professional: at Bosch (2020-2022) Linux servers carried everything - ML platform, Docker hosts, backend services, Ansible control - and he administered and automated them (see BS-6). Personal and current: a self-hosted Debian server he has run and hardened for ~10 years (see PP-3). Hardening and security configuration evidence leans on the personal side; say so rather than implying an employer mandate. Does NOT license claiming datacenter/physical infrastructure work, which remains a genuine gap."
},
{
"name": "Docker",
"evidence": "production-current-and-historical",
"output": "allowed"
},
{
"name": "Java",
"evidence": "production-historical",
"output": "allowed-with-context"
},
{
"name": "C#",
"evidence": "production-historical",
"output": "allowed-with-context"
},
{
"name": "C++",
"evidence": "production-historical-limited",
"output": "allowed-with-context"
},
{
"name": "JavaScript",
"evidence": "production-historical-limited",
"output": "allowed-with-context"
},
{
"name": "LiteLLM",
"evidence": "hands-on-current",
"output": "allowed-with-context"
},
{
"name": "custom GPTs",
"evidence": "hands-on-current",
"output": "allowed-with-context"
},
{
"name": "Kiro",
"evidence": "hands-on-current",
"output": "allowed-with-context"
},
{
"name": "Copilot",
"evidence": "hands-on-current",
"output": "allowed-with-context"
},
{
"name": "quantitative/systematic trading concepts",
"evidence": "coursework-and-personal-project",
"output": "allowed-with-context",
"note": "Udacity AI for Trading nanodegree (PP-2) plus the self-built momentum platform (PP-1): factor construction, cross-sectional/residual momentum, ATR stops, backtesting. Self-directed only — NO professional quant, trading or finance experience. Never list under a Professional Experience skills line without the personal-project context, and never as a quant qualification."
},
{
"name": "backtesting",
"evidence": "personal-project",
"output": "allowed-with-context",
"note": "PP-1 scheduled backtest pipelines, personal scale. Never quote a result or performance figure."
},
{
"name": "TensorFlow/Keras",
"evidence": "certification",
"output": "certification-context-only"
},
{
"name": "PyTorch",
"evidence": "certification",
"output": "certification-context-only",
"note": "IBM AI Engineering Professional Certificate (Coursera, 4 Jun 2020) includes the course \"Deep Neural Networks with PyTorch\"; certificate PDF verified 2026-08-27. Same footing as TensorFlow/Keras: certification context only, never presented as production or project experience."
},
{
"name": "TypeScript",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "FastAPI",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "Flask",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "Django",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "LangChain",
"evidence": "never-used",
"output": "forbidden"
},
{
"name": "LangGraph",
"evidence": "never-used",
"output": "forbidden"
},
{
"name": "LlamaIndex",
"evidence": "never-used",
"output": "forbidden"
},
{
"name": "formal model evaluation",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "LLM fine-tuning",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "Azure",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "GCP",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "Terraform",
"evidence": "unverified",
"output": "forbidden"
},
{
"name": "GitOps",
"evidence": "pending-user-clarification",
"output": "forbidden",
"note": "RESOLVED 2026-08-27, and the answer is NO. User confirmed: Ansible playbooks lived in Git, but execution was PUSH-based out of Jenkins/GitLab pipelines - there was no pull-based reconciliation agent (Argo CD, Flux). That is version-controlled, CI-driven infrastructure automation, not GitOps. The word stays forbidden in every document. The SUBSTANCE is writable and should be written: 'versionskontrollierte Infrastrukturautomatisierung mit Ansible aus Git, ausgefuehrt ueber Jenkins-/GitLab-Pipelines'. The Kdo Cy JD names 'Pull-GitOps' explicitly, so precision here is an asset, not a loss."
}
],
"global_forbidden_output_patterns": [
"petabyte scale",
resume_builder/experience/experience_bundeswehr.md
+5
View File
@@ -9,14 +9,19 @@
### Achievement BW-1: Officer Training and Service
**Canonical ID:** BW-1
**Last verified:** 2026-08-27
**User's role:** Officer candidate / officer
**Status:** Completed service; resigned as Second Lieutenant
**Ownership scope:** Individual service and completed training. No command scope, deployment or technical military responsibility is established.
**Verified result / metric state:** Six years of service and separation as Second Lieutenant are verified; no performance metric is established.
**Context:** Completed officer candidate training and officer school, then served in the German Armed Forces for six years.
**Safe bullet direction:** Completed officer candidate training and officer school during six years in the German Armed Forces (Bundeswehr), leaving service as Second Lieutenant. Do not imply a current clearance, NATO service, combat role, technical AI work, or responsibilities not verified.
**ATS keywords:** German Armed Forces, Bundeswehr, officer, multinational environment, operational context, structured leadership
**Relevant role context:** Defence, military and public-security employers; organisational familiarity rather than technical evidence.
**Reframing notes:**
- Defence / NATO roles: Include as concise context for operational judgement and organisational familiarity.
resume_builder/helpers/validate_resume_system.py
+62
View File
@@ -137,6 +137,67 @@ def scan_document(path: Path, claims: dict) -> list[str]:
return errors
TRIPLE = re.compile(r"\w+, [^,]{3,45}, [^,]{3,45} and |\w+, [^,]{3,45} and ")
def _bullet_blocks(text: str) -> list[list[str]]:
"""Bullets grouped by rSubsection (one group per position)."""
body = re.search(r"\\begin\{document\}(.*)\\end\{document\}", text, re.S)
body_text = body.group(1) if body else text
blocks = re.split(r"\\begin\{rSubsection\}", body_text)[1:]
return [re.findall(r"\\item\s+(.+)", block) for block in blocks]
def _visible_text(bullet: str) -> str:
"""Bullet text with LaTeX markup removed, so \\textbf{Owned ...} reads as 'Owned ...'."""
text = re.sub(r"\\[a-zA-Z]+\*?(\[[^]]*\])?", " ", bullet)
return re.sub(r"[{}$\\]", " ", text)
def _opening_word(bullet: str) -> str:
match = re.search(r"[A-Za-z]+", _visible_text(bullet))
return match.group(0).lower() if match else ""
def cadence_checks(path: Path) -> list[str]:
"""Bullet cadence diagnostics (resume_reference.md 7a).
Warnings only, never errors. Uniform rhythm is a style flaw, not a truth flaw:
no claim, scope or hedged verb may be bent to satisfy these. Reshape sentences.
"""
blocks = _bullet_blocks(read_text(path))
bullets = [b for block in blocks for b in block]
if len(bullets) < 6:
return []
warnings: list[str] = []
triples = [b for b in bullets if TRIPLE.search(b)]
share = len(triples) / len(bullets)
if share > 0.5:
warnings.append(
f"{path}: {len(triples)}/{len(bullets)} bullets ({share:.0%}) use an "
f"'X, Y and Z' triple (reference 7a: under ~50%). Reshape sentences, not claims."
)
for position, block in enumerate(blocks, 1):
for index in range(1, len(block)):
previous, current = _opening_word(block[index - 1]), _opening_word(block[index])
if previous and previous == current:
warnings.append(
f"{path}: position {position}, bullets {index}-{index + 1} both open "
f"with '{previous}'"
)
# Threshold calibrated on the 50 multi-bullet positions in output/: spreads are
# bimodal (1-9, then 15-17). <=4 flags the tight tail (~28%), not the median.
lengths = [len(_visible_text(b).split()) for b in block]
if len(lengths) >= 4 and max(lengths) - min(lengths) <= 4:
warnings.append(
f"{path}: position {position} bullets cluster at {min(lengths)}-{max(lengths)} "
f"words; vary length within the position, not just across the document."
)
return warnings
def main() -> int:
parser = argparse.ArgumentParser()
parser.add_argument("--document", action="append", default=[], help="Generated .tex file to validate")
@@ -154,6 +215,7 @@ def main() -> int:
errors.append(f"document not found: {path}")
else:
errors.extend(scan_document(path, claims))
warnings.extend(cadence_checks(path))
for warning in warnings:
print(f"WARN: {warning}")
resume_builder/reference/critical_rules.md
+3
View File
@@ -7,6 +7,9 @@
5. Put employer, formal title and dates before any narrative framing.
6. Use an optional 2--3 line summary, 4--6 skills lines and normally 11--14 evidence-led bullets.
7. Mix natural bullet lengths. There is no character target and no page-fill quota.
7a. Vary bullet cadence: keep "X, Y and Z" triples under about half the bullets, avoid two
adjacent bullets opening with the same verb, and vary length *within* each position.
Style only — never bend a claim, a scope or a hedged verb to fix rhythm (see §7a).
8. Do not list unverified skills, metrics, customers, scale or causal impact.
9. Scope Swisscom migration and Data Mesh claims to Dennis's domains, components and products.
10. Security Champion means the 2025/2026 team role only; omit by default.
resume_builder/reference/critique_framework.md
+1 -1
View File
@@ -42,7 +42,7 @@ Score the candidate-role pairing before reading tailored prose.
| Bullet evidence and impact | 20 | Are bullets specific and scoped without invention? |
| Relevance and terminology | 15 | Does it cover the JD naturally? |
| Skills evidence | 10 | Is every skill evidence-backed and interview-ready? |
| Mechanics/readability | 10 | Does it compile, parse and read cleanly? |
| Mechanics/readability | 10 | Does it compile, parse and read cleanly? Includes **cadence variety** — read the bullets in sequence and count "X, Y and Z" triples, repeated opening verbs and length clustering per position (`resume_reference.md` §7a). Deduct at most 1 pt; it is a style flaw, never a truth flaw. |
Truth/provenance below 8/10 is an automatic document failure.
resume_builder/reference/resume_reference.md
+32
View File
@@ -96,6 +96,36 @@ Natural one-, two- and three-line bullets may be mixed. There are no target char
Use metrics only when verified. Company size, customer names, generic industry volumes and market economics are context, not personal impact.
### 7a. Cadence variety (anti-monotony)
Individually honest bullets can still read as machine-written when they all share one
rhythm. Measured across the 18 packages in `output/` (368 bullets): **45 % used the same
"X, Y and Z" triple**, and the SBB package reached **85 % — 11 of 13 bullets, every
Swisscom and Bosch line**. Nothing in it was false; it simply had one cadence.
This is a *style* problem, never a truth problem. Never trade accuracy, scope discipline
or a hedged verb to fix cadence — an accurate monotonous bullet beats a varied inaccurate
one. Fix it by reshaping sentences, not by changing what is claimed.
Per document, aim for:
- **At most about half the bullets carrying a three-part list.** The triple is normal resume
grammar and humans use it too; the tell is using little else.
- **At least two or three bullets in a different shape** — a short declarative with a single
object; a bullet with one object rather than three; a bullet whose scope clause comes first.
- **No two consecutive bullets opening with the same verb** (the SBB resume had "Build and
model…" directly followed by "Build and operate…").
- **Visible length variation inside each position**, not just across the document. Five bullets
at 25–30 words read as generated even when the document's overall range looks fine.
`validate_resume_system.py --document <file.tex>` measures all three and reports them as
**WARN**, never ERROR — cadence can never fail a document. The length threshold is calibrated
on the 50 multi-bullet positions in `output/`, whose spreads are bimodal (1–9 words, then
15–17): it flags a spread of 4 words or less, the tight tail, not the median.
Diagnostic only — like character counts, these are checks, never targets, and no bullet
should be padded or trimmed to hit them.
## 8. Titles and Seniority
- Preserve official titles when recognizable.
@@ -113,6 +143,8 @@ After generation:
4. Extract text with `pdftotext` and confirm employer/title/date order.
5. Inspect the rendered PDF: no clipping, overlap, tiny text, isolated headings or awkward page break.
6. Check that experience begins comfortably on page 1 and certifications are not duplicated.
7. Act on any cadence WARN from step 1 (§7a). These never fail the document; reshape the
offending sentences, never the claims behind them.
Do not add content merely to reduce bottom whitespace.