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 Withdrawn.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

CLAUDE.md

@@ -157,7 +157,7 @@ _Update this section when starting/finishing a JD._
 
 | Session | Status | Next Command |
 |---------|--------|-------------|
-| Citadel Securities — Platform Engineer, Research Platform Engineering, **Zürich** (also NY/Miami) | **PHASE 0 DONE 2026-08-25 — awaiting user decision, Phase 1 not started.** Evidence Fit **74/100 (Adjacent, top of band, 1 pt below Core)**, hard gate **PASS with a flagged risk**. **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. | User decision: `/make-resume` Phase 1 to proceed, or record NO-GO. If proceeding, seek a warm channel first — cold is the one variable that has never worked |
+| Citadel Securities — Platform Engineer, Research Platform Engineering, **Zürich** (also NY/Miami) | **PHASE 0 + PHASE 1 PLAN DONE — BLOCKED 2026-08-26 on the working-model question.** 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) | **SUBMITTED 2026-08-25.** Cohort slot: **Core (evidence-first-2026-01)**. 2pp English resume (13 bullets, validator PASS) + critique: **Document Quality 92/100** (90 at critique, +2 after Edit 1), hard gate PASS, no Tier 1 truth findings, **no outstanding fixes**. No cover letter — SBB waives it. Evidence Fit **79/100 (Core, lower end)**, hard gate **PASS** — the title-defining capability (build/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 + B permit**. Real strength cluster for the *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/comp not fit:** (1) **Anforderungsniveau K** is a capped GAV band and the figures are not public — must be asked, not researched; (2) the seat reads **lateral or below** current Staff + Component Owner scope (same shape as the declined BKW). Named gaps: Snowflake, dbt, Argo Workflows, Helm, Power BI, Spring Boot, Angular — all non-canonical, survivable only because the JD prefixes the stack list with "z. B.". **Cover letter: NO — SBB explicitly waives it.** Channel: posting names **Andri Wienandts, People Leader, +41 79 364 62 53** — the first published warm entry point in the entire log. **The comp question was NOT resolved before submitting**, so the K band is now a live screening topic rather than a pre-cleared one; the same applies to Kidz Care childcare support (up to 90% of Betreuungskosten but scaled on *gross household* income, so realistically far less at this level — bracket table is intranet-only). | Prep an interview brief before any screening call: (1) Anforderungsniveau K band figures, (2) design/architecture scope vs current Staff + Component Owner level, (3) Kidz Care actual rate at this band. Named gaps to have honest answers for: Snowflake, dbt, Argo Workflows, Helm, Power BI |
 | 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 |

output/CitadelSecurities_PlatformEngineer/session_citadel_securities_platform_engineer.md

@@ -15,9 +15,9 @@
 ## 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: **74/100**
+- Evidence Fit: **71/100** (revised down from 74 on 2026-08-26 — see Working Model below)
-- Fit class: **Adjacent (top of band, one point below Core)**
+- Fit class: **Adjacent (mid-band)**
-- Hard gate: **PASS, with one flagged risk** — see Gap Assessment. No minimum qualification is a Gap; nothing required is `forbidden` in `claims.json`.
+- 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.
@@ -85,6 +85,38 @@
 
 **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.
@@ -124,10 +156,92 @@
 - 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.
+
 ## Status
 
 - Phase 0: **DONE** (2026-08-25)
-- Phase 1: **NOT STARTED** — user requested Phase 0 only
+- Phase 1: **PROPOSED 2026-08-26 — awaiting user confirmation**
 - Resume: PENDING
 - Cover Letter: PENDING (decision YES)
 - Critique: PENDING