feat(app): NVIDIA Senior Production Engineer, DGX Cloud submitted; on-duty, GPU and networking KB gaps closed; triage 2026-10-22

Application (submitted 2026-10-22, Workday):
- NVIDIA JR2027105, Switzerland Remote. Evidence Fit 65, Stretch (SLIs and
  error budgets not evidenced), hard gate PASS, channel Weak, no cover letter.
- Package in output/NVIDIA_ProdEng_DGX_Cloud/: session, resume .tex, critique
  (Document Quality 85 before Edit 1; Tier 1 + all Tier 2 applied), verbatim JD.
- Prior NVIDIA req JR2024256 was pulled (last listed 2026-09-18; page gone
  2026-10-22): noted on that application, so there was no stacking conflict.

KB (claims.json, user-confirmed 2026-10-22):
- SW-11: current rotating engineer-on-duty role, Opsgenie alerting, alerts and
  bugs auto-create Jira tickets; he writes alert thresholds (e.g. data quality).
- BS-1: the Bosch inference ran on NVIDIA GPUs; CUDA registered forbidden as a
  skill (system context only, little configuration done).
- BS-4: the PoC included Grafana, Prometheus and Loki; ELK Stack, Grafana,
  Prometheus and Loki registered as PoC-context skills.
- networking: certification context only (AWS SA Associate); work uses
  automated landing zones.
- Extraction and CLAUDE.md log row added.

Scout state:
- decisions: RUAG Senior DevOps (Kubernetes) C5I parked (home-office terms);
  SLB Data Scientist skipped (user); NVIDIA JR2008695 skipped (C++/Rust 8y);
  JR2027105 applied; Norway/Apple/Swiss JD-read triage decisions.
- employer_reputation: SLB (Glassdoor 3.9), MazeMap and SpareBank 1
  Nord-Norge (unchecked).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CLAUDE.md

@@ -183,8 +183,9 @@ _See `config.md` for user-specific corrections. Add verified errors here as you
 
 | Date | Correction | Files fixed |
 |------|-----------|-------------|
+| 2026-10-22 | **Current on-call was unrecorded, which made the KB look as if his production-operations evidence ended at ART B2B. This is the eighth gap of the recurring ops shape, surfaced by the NVIDIA *Senior Production Engineer, DGX Cloud* (JR2027105) fit check.** User-confirmed in his current team: **Opsgenie** alerting, a **rotating engineer-on-duty role** he takes part in, and bugs and alerts that automatically create Jira tickets the team assigns and fixes. Recorded as **SW-11**. A follow-up confirmed that **he writes the alert thresholds for his own pipelines**, data-quality thresholds being the named example. The Opsgenie/Jira integration is not his and "24/7" is not established. The follow-up also confirmed that the Bosch BS-4 PoC included **Grafana, Prometheus and Loki** (previously only in his earlier CV); registered as PoC-context skills with ELK Stack. **Same message:** (a) the Bosch ML inference (BS-1) **ran on NVIDIA GPUs with CUDA**, but he "never did much configuration there". The writable form is "GPU-backed inference on NVIDIA GPUs". **CUDA is registered `forbidden`** as a skill (it is system context only), and GPU tuning is forbidden on BS-1. (b) **Networking: no professional configuration.** Day-to-day work runs on automated AWS landing zones, and AWS networking is covered only by the Solutions Architect – Associate certificate. Recorded as `networking`, certification context only. **Lesson, the same as before: on an SRE/production JD, ask about the CURRENT on-call and alerting setup before scoring. The past SW-2 Rufbereitschaft had hidden the current one.** | `claims.json` (SW-11, BS-1, skills Opsgenie/CUDA/networking); `knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md` |
 | 2026-10-21 | **Software design and architecture was under-recorded: the seventh canonical gap of the recurring shape, surfaced by the Google *Staff SWE, Shopping Graph* minimum *"3 years of experience with software design and architecture."*** Before this, the only design evidence was SW-7 (current, undated, 2026-10-03) and a *contribution* at Generali (GN-3), so the line read as unproven. User-confirmed: **software design and architecture were part of his work at Bosch (ingestion and inference scripts, 2020-2022) and at Swisscom (data products)**, and both count toward the year requirement; together they exceed three years. Recorded on BS-1, BS-2 and SW-7, with `designed` added to the BS-1/BS-2 allowed verbs. **Scope kept tight:** at Bosch, "designed" covers the ingestion scripts and the inference integration only, not the data services in general (the first patch over-extended it and was trimmed the same session). Design never means an architect title, design authority or shared-platform architecture. The Swisscom start date stays undated. **Lesson, the same as 10-08 and 10-14: when a JD has a years-of-X minimum that the canonical file states without a duration, ask for the duration before scoring.** | `claims.json` (BS-1, BS-2, SW-7); `config.md` (data-product design row); `knowledge_base/extractions/2026-10-21_software_design_architecture_clarification.md` |
-| 2026-10-22 | **Scout findings fixed in one pass; the biggest one is that Apple had been blind to Switzerland for about two months.** (a) **Apple:** the `location=switzerland-CHE` slug stopped resolving and Apple silently applied a US location ("Chesterfield, Missouri"). Every run from at least 2026-08-31 to 2026-10-22 returned the same 21 US/intern cards and **0 Swiss roles**, which read as "Apple has nothing" rather than as a failure. The code is now `CHEC` (from `/api/v1/refData/postlocation`); the card is the result list item, so the location is readable; and there is pagination, a settle delay and per-element timeouts (one run took 876s, the next 25s). Result: **24 Swiss roles**, Zurich AIML/infra included. (b) **BIS:** bis.org went behind a Cloudflare challenge (403, even for a headless browser), so the board moved to its Workday tenant `bis.wd3.myworkdayjobs.com/External`, which also gives the real office. A host alias plus a lower-cased jr-id keep all 9 `bis.org#jr…` decisions linked (applied jr100498 included). (c) **NAV:** the search API 429s after ~9 requests and then for hours. Discovery now stays at 8 searches, and posting text comes from NAV's official third-party feed (`pam-stilling-feed`, public rotating token), which tolerated bursts. The feed lacks ~2/3 of ads, so a small, slow website-page fallback covers those, and verdicts are cached. (d) **EMEA leak fixed:** a country allowlist without CH/DK plus a trailing "EMEA" tag is no longer Europe-remote (27 rows removed). Swiss/Danish segments in multi-location strings are checked per segment, so "PL-Poland-Remote \| CH-Zurich-MSO" stays Swiss. A regression over all 2026-10-22 rows showed only the fenced rows changing. (e) **Reputation:** every board now has a dated verdict. Glassdoor and **kununu** are both primary sources, and whichever has more reviews decides (user tip: kununu is far better for Swiss employers). AVOID: **BIS** (Glassdoor 2.9/56, where an application is pending), **BFH** (kununu 3.3/132), **Kdo Cy** (federal-administration kununu 3.4/997, not unit-specific). **Lesson: a board that returns the same non-zero count with zero eligible rows for weeks is a silent failure, not an empty market. Check what filter the site actually applied.** | `scout.py` (Apple, BIS, NAV, location filter, reputation); `state/employer_reputation.json` |
+| 2026-10-22 | **Scout findings fixed in one pass; the biggest one is that Apple had been blind to Switzerland for about two months.** (a) **Apple:** the `location=switzerland-CHE` slug stopped resolving and Apple silently applied a US location ("Chesterfield, Missouri"). Every run from at least 2026-08-31 to 2026-10-22 returned the same 21 US/intern cards and **0 Swiss roles**, which read as "Apple has nothing" rather than as a failure. The code is now `CHEC` (from `/api/v1/refData/postlocation`); the card is the result list item, so the location is readable; and there is pagination, a settle delay and per-element timeouts (one run took 876s, the next 25s). Result: **24 Swiss roles**, Zurich AIML/infra included. (b) **BIS:** bis.org went behind a Cloudflare challenge (403, even for a headless browser), so the board moved to its Workday tenant `bis.wd3.myworkdayjobs.com/External`, which also gives the real office. A host alias plus a lower-cased jr-id keep all 9 `bis.org#jr…` decisions linked (applied jr100498 included). (c) **NAV:** the search API 429s after ~9 requests and then for hours. Discovery now stays at 8 searches, and posting text comes from NAV's official third-party feed (`pam-stilling-feed`, public rotating token), which tolerated bursts. The feed lacks ~2/3 of ads, so a small, slow website-page fallback covers those, and verdicts are cached. (d) **EMEA leak fixed:** a country allowlist without CH/DK plus a trailing "EMEA" tag is no longer Europe-remote (27 rows removed). Swiss/Danish segments in multi-location strings are checked per segment, so "PL-Poland-Remote \| CH-Zurich-MSO" stays Swiss. A regression over all *eligible* 2026-10-22 rows showed only the fenced rows changing (rows newly admitted by the per-segment check were not regression-tested). (e) **Reputation:** every board now has a dated verdict. Glassdoor and **kununu** are both primary sources, and whichever has more reviews decides (user tip: kununu is far better for Swiss employers). AVOID: **BIS** (Glassdoor 2.9/56, where an application is pending), **BFH** (kununu 3.3/132), **Kdo Cy** (federal-administration kununu 3.4/997, not unit-specific). **Lesson: a board that returns the same non-zero count with zero eligible rows for weeks is a silent failure, not an empty market. Check what filter the site actually applied.** | `scout.py` (Apple, BIS, NAV, location filter, reputation); `state/employer_reputation.json` |
 | 2026-10-21 | **FIXED 2026-10-22 (Microsoft links; the EMEA leak below was fixed the same day, see the 2026-10-22 row).** Fix: `fetch_pcsx` now emits `apply.careers.microsoft.com<positionUrl>?jobId=<displayJobId>`, a `?jobId=` pattern precedes the `/job/` one in `_JOB_ID_PATTERNS`, and `_HOST_ALIASES` maps the new host to the old so `_decision_key` is unchanged. Verified: all 7 logged Microsoft decisions still resolve (applied 200043897 included), and a new-form link opened the real posting with the job number on the page. Reports before 2026-10-22 still carry the dead form. Original finding: **every Microsoft link the scout emits is dead, in every report**, found in the 2026-10-20 triage. `fetch_pcsx` builds `jobs.careers.microsoft.com/global/en/job/<displayJobId>`, which now redirects to a generic `apply.careers.microsoft.com/careers?...&pid=1970393557041551` search page. Verified on 200060867 and on the **pending application 200043897** (Principal FDE). The application itself is unaffected; only the link is. The working form is `https://apply.careers.microsoft.com/careers/job/<pcsx id>`, the API's `positionUrl`. **Deliberately NOT patched yet:** 6 logged Microsoft decisions (including 200043897) are keyed on `displayJobId` through the `careers\.microsoft\.com /job/(\d+)` pattern. Swapping the URL would re-key them on the pcsx id and orphan them, the same failure as the 09-04 RUAG/BIS de-linking. **Safe fixes:** (a) keep `url` and add a working `apply_url` field to the row, or (b) emit the new URL with the display id carried in it, plus a host-normalised key pattern. Either way, verify all 6 rows still resolve through `find_decision_url` afterwards. Same triage, also OPEN: an **"EMEA" region suffix lets country-fenced roles through the location filter** (GitLab FR/DE/NL/ES/UK, Grafana ×5, Fivetran Serbia: 7 of 109 new rows), which breaks the docstring rule against multi-country allowlists that omit CH. | not yet fixed: `scout.py` (`fetch_pcsx`, location filter); evidence in `reports/2026-10-20.triage.md` |
 | 2026-10-15 | **Two scout adapters were reporting wrong data without failing; both were found by the 2026-10-14 triage.** (a) **BIS:** the RSS adapter stamped `default_location: Basel, Switzerland` on every vacancy, so **jr100491 (Hong Kong, BIS Asia office)** passed the CH filter and reached a JD read. The feed carries no location, but each vacancy page has a static *Geographical location* field. `fetch_rss` now reads it per item (`location_page_regex`), falls back to a *"based in ..."* sentence in the feed description, and only then uses `default_location`, printing a warning when it does. Verified: Basel reqs still pass, and jr100491 now reads *8 Finance Street, Central, Hong Kong* and is excluded. (b) **Confluent:** IBM closed the acquisition on 2026-03-17, and the Ashby board went from 17 jobs (10-08) to `{"jobs":[]}` (10-14). careers.confluent.io sits behind a Vercel bot checkpoint. The roles are now IBM requisitions, served by the open IBM careers search API (`www-api.ibm.com/search/api/v2`, POST, `size` capped at 100, pages with `from`). New `ibm` adapter; Confluent keeps its id, so seen-state and decisions still link. First run: 84 roles, 0 in CH/Europe-remote scope. **HashiCorp was dropped on 2026-06-06 for having "no clean public ATS API" after IBM absorbed it. That API exists, so HashiCorp can come back through the same adapter if wanted.** **Rule: when an adapter must assume a field, the assumption has to be visible. A silent default looks exactly like real data.** | `scout.py` (`fetch_rss`, `_rss_item_location`, `fetch_ibm`, BIS and Confluent config) |
 | 2026-10-14 | **Spotfire was the sixth evidenced-but-unregistered skill, and junior mentoring was unrecorded - both surfaced by the BIS Basel jr100498 *Senior/Lead Data & Analytics Engineer* (MED IT) triage.** (a) **TIBCO Spotfire** sat in the BS-5 and BS-8 scopes (co-owned environment, technical Application Owner, C# extensions, TAF 2022, spatial wafer-data map charts) but not in the `skills` array, so the JD line *"using database reporting and business intelligence tools"* had no citable skill; registered `production-historical`, `allowed` on user instruction. It does not license Power BI, Tableau or Qlik. (b) The JD requires *"Experience providing general supervision and task-specific guidance to junior colleagues"*, and the only canonical evidence was BS-7 (students, one thesis as Zweitpruefer). User-confirmed: he **onboarded and mentored junior developers at Bosch and at Swisscom** - recorded as **BS-9** and **SW-10**, one per employer. The Swisscom phase (ART B2B or ART DPT) is not recorded, so write it in past tense. Neither is people management. (c) **SDMX: user-confirmed never used**; registered `forbidden` so the scanner catches it. It is an accountability line in the JD, not a qualification. (d) The same audit found **Teradata, Oracle Database and Hadoop/Impala** evidenced only in the SW-2/BS-2/BS-8 scopes and absent from the `skills` array, although every past skills line listed them; registered `production-historical` from those scopes (Oracle explicitly does not establish DBA work). **Lesson, same as 10-08: when a JD has a supervision or mentoring line, ask about junior colleagues before falling back on BS-7. Student supervision is weaker evidence and the stronger evidence was simply unrecorded.** | `claims.json` (BS-9, SW-10; skills TIBCO Spotfire, SDMX, Teradata, Oracle Database, Hadoop / Impala); `knowledge_base/extractions/2026-10-14_mentoring_spotfire_sdmx_clarification.md` |

job_scout/state/application_cohort.json

@@ -26,7 +26,7 @@
       "outcome": "applied",
       "id": "app-9068f462d9b9",
       "label": "NVIDIA - Senior DevOps Engineer, Zürich (req JR2024256)",
-      "next_action": "Await response. At first contact, clarify homeoffice/onsite requirements and total package. Baseline is 3 homeoffice days/week; fewer days require a clear overall improvement. Note: `claims.json` has no shell-scripting entry, so **Bash is omitted** from Skills although the JD names it and ten years of Debian admin plus Ansible work plainly involve it - same shape as the Ansible/Linux gaps found 2026-08-27. Confirm with the user and record it. Groovy stays out regardless.",
+      "next_action": "No action expected; req pulled ~2026-09-20. Treat as dormant. Does not block NVIDIA JR2027105 (Senior Production Engineer, DGX Cloud).",
       "current_summary": "SUBMITTED 2026-09-05.",
       "events": [
         {
@@ -39,6 +39,11 @@
         {
           "type": "applied",
           "date": "2026-09-05"
+        },
+        {
+          "type": "note",
+          "date": "2026-10-22",
+          "note": "Posting pulled: JR2024256 was last listed in the 2026-09-18 scout scrape, absent from every scrape since 2026-09-23, and the posting page reads 'The page you are looking for doesn't exist' on 2026-10-22. No response from NVIDIA recorded. Outcome left as applied (no rejection received). The req is closed, so it no longer blocks a second NVIDIA application (BIS jr100429 precedent)."
         }
       ],
       "cohort_id": "evidence-first-2026-01"
@@ -475,6 +480,31 @@
           "date": "2026-10-21"
         }
       ]
+    },
+    {
+      "id": "app-b13ee50001fc",
+      "company": "NVIDIA",
+      "role": "Senior Production Engineer, DGX Cloud, Switzerland Remote (JR2027105)",
+      "label": "NVIDIA — Senior Production Engineer, DGX Cloud, Switzerland Remote (JR2027105)",
+      "application_date": "2026-10-22",
+      "fit_class": "stretch",
+      "evidence_fit": 65.0,
+      "hard_gate": "pass",
+      "channel": "weak",
+      "outcome": "applied",
+      "cohort_id": "evidence-first-2026-01",
+      "job_id": "JR2027105",
+      "url": "https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105",
+      "fit_override_reason": "Stretch: required SRE line (SLIs, SLOs, error budgets) only Adjacent; hard gate PASS. User chose it 2026-10-22 over the alternatives (remote CH, NVIDIA rating 4.4); prior NVIDIA req JR2024256 pulled, so no stacking.",
+      "rubric_version": "critique-framework-2026-09-11",
+      "next_action": "Await response. Send the LinkedIn note to an NVIDIA recruiter or DGX Cloud engineer (draft in session). Interview prep: incident debugged end to end (SW-2/SW-11); safe model rollout in the fab (BS-1, rollout mechanics not in KB); Kubernetes recency; distributed systems at scale; vLLM/SGLang/Dynamo not used. Package: output/NVIDIA_ProdEng_DGX_Cloud/.",
+      "current_summary": "applied",
+      "events": [
+        {
+          "type": "applied",
+          "date": "2026-10-22"
+        }
+      ]
     }
   ]
 }

job_scout/state/applications.json

@@ -28,7 +28,7 @@
       "outcome": "applied",
       "id": "app-9068f462d9b9",
       "label": "NVIDIA - Senior DevOps Engineer, Zürich (req JR2024256)",
-      "next_action": "Await response. At first contact, clarify homeoffice/onsite requirements and total package. Baseline is 3 homeoffice days/week; fewer days require a clear overall improvement. Note: `claims.json` has no shell-scripting entry, so **Bash is omitted** from Skills although the JD names it and ten years of Debian admin plus Ansible work plainly involve it - same shape as the Ansible/Linux gaps found 2026-08-27. Confirm with the user and record it. Groovy stays out regardless.",
+      "next_action": "No action expected; req pulled ~2026-09-20. Treat as dormant. Does not block NVIDIA JR2027105 (Senior Production Engineer, DGX Cloud).",
       "current_summary": "SUBMITTED 2026-09-05.",
       "events": [
         {
@@ -41,6 +41,11 @@
         {
           "type": "applied",
           "date": "2026-09-05"
+        },
+        {
+          "type": "note",
+          "date": "2026-10-22",
+          "note": "Posting pulled: JR2024256 was last listed in the 2026-09-18 scout scrape, absent from every scrape since 2026-09-23, and the posting page reads 'The page you are looking for doesn't exist' on 2026-10-22. No response from NVIDIA recorded. Outcome left as applied (no rejection received). The req is closed, so it no longer blocks a second NVIDIA application (BIS jr100429 precedent)."
         }
       ],
       "cohort_id": "evidence-first-2026-01"
@@ -863,6 +868,31 @@
           "date": "2026-10-21"
         }
       ]
+    },
+    {
+      "id": "app-b13ee50001fc",
+      "company": "NVIDIA",
+      "role": "Senior Production Engineer, DGX Cloud, Switzerland Remote (JR2027105)",
+      "label": "NVIDIA — Senior Production Engineer, DGX Cloud, Switzerland Remote (JR2027105)",
+      "application_date": "2026-10-22",
+      "fit_class": "stretch",
+      "evidence_fit": 65.0,
+      "hard_gate": "pass",
+      "channel": "weak",
+      "outcome": "applied",
+      "cohort_id": "evidence-first-2026-01",
+      "job_id": "JR2027105",
+      "url": "https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105",
+      "fit_override_reason": "Stretch: required SRE line (SLIs, SLOs, error budgets) only Adjacent; hard gate PASS. User chose it 2026-10-22 over the alternatives (remote CH, NVIDIA rating 4.4); prior NVIDIA req JR2024256 pulled, so no stacking.",
+      "rubric_version": "critique-framework-2026-09-11",
+      "next_action": "Await response. Send the LinkedIn note to an NVIDIA recruiter or DGX Cloud engineer (draft in session). Interview prep: incident debugged end to end (SW-2/SW-11); safe model rollout in the fab (BS-1, rollout mechanics not in KB); Kubernetes recency; distributed systems at scale; vLLM/SGLang/Dynamo not used. Package: output/NVIDIA_ProdEng_DGX_Cloud/.",
+      "current_summary": "applied",
+      "events": [
+        {
+          "type": "applied",
+          "date": "2026-10-22"
+        }
+      ]
     }
   ],
   "targeting_notes": [

job_scout/state/decisions.json

@@ -1126,13 +1126,6 @@
     "note": "Claude judgment 2026-07-06: GPU networking research specialization",
     "date": "2026-07-06"
   },
-  "https://nvidia.wd5.myworkdayjobs.com/job/Switzerland-Zurich/Senior-Software-Developer_JR2008695": {
-    "company": "NVIDIA",
-    "title": "Senior Software Developer",
-    "decision": "maybe",
-    "note": "Claude judgment 2026-07-06: generic title, no domain signal from the title alone - worth a quick JD read",
-    "date": "2026-07-06"
-  },
   "https://nvidia.wd5.myworkdayjobs.com/job/Switzerland-Zurich/Senior-Software-Developer--AI-Networking_JR2016186": {
     "company": "NVIDIA",
     "title": "Senior Software Developer, AI Networking",
@@ -1504,13 +1497,6 @@
     "note": "Claude judgment 2026-07-06: Senior SWE with practical AI/ML experience (not research-depth), local tier",
     "date": "2026-07-06"
   },
-  "https://jobs.ruag.ch/offene-stellen/software-engineer-c5i/ec08db9a-3b4f-4aad-8eb5-6134c5e87458": {
-    "company": "RUAG (Thun/Bern)",
-    "title": "Software Engineer C5I m/f/d",
-    "decision": "maybe",
-    "note": "Claude judgment 2026-07-06: same C5I domain as the DevOps req but a non-senior/generic SWE variant",
-    "date": "2026-07-06"
-  },
   "https://careers.roche.com/global/en/job/202603-108151/Sr-Scientist-Computational-Biology-Neuroscience-Rare-Diseases": {
     "company": "Roche",
     "title": "Sr. Scientist - Computational Biology - Neuroscience & Rare Diseases",
@@ -1553,13 +1539,6 @@
     "note": "Claude judgment 2026-07-06: internship",
     "date": "2026-07-06"
   },
-  "https://career2.successfactors.eu/sfcareer/jobreqcareer?jobId=1476&locale=en_US&company=Swissgrid": {
-    "company": "Swissgrid (Aarau)",
-    "title": "(Senior) Cloud Security Engineer (80-100%)",
-    "decision": "maybe",
-    "note": "Claude judgment 2026-07-06: cloud is a core skill, security specialization a partial gap, local tier",
-    "date": "2026-07-06"
-  },
   "https://career2.successfactors.eu/sfcareer/jobreqcareer?jobId=912&locale=de_DE&company=Swissgrid": {
     "company": "Swissgrid (Aarau)",
     "title": "Bachelor's- & Master's thesis",
@@ -1817,13 +1796,6 @@
     "note": "REJECTED 2026-09-10, 14 days after the 2026-08-27 submission, no interview. Form letter from Besarta Mehmeti with no diagnostic content. Was Core, Evidence Fit 79, hard gate PASS, Document Quality 92, channel weak. Marcel Matthey-Doret's published line was never called. Kdo Cy had 11 roles open on the portal at submission time, so the ORGANISATION is not closed - only this req. 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": "2026-09-10"
   },
-  "https://jobs.ruag.ch/offene-stellen/senior-devops-kubernetes-engineer-c5i/2eef27f6-9546-499d-8492-c1cec22f548a": {
-    "company": "",
-    "title": "",
-    "decision": "shortlist",
-    "note": "Rank 1 and the only Core-classifiable target in the log. Re-verified live on the board 2026-09-04. Every required line has canonical evidence: Kubernetes SW-3/SW-7, Linux BS-6/PP-3, GitLab CI-CD SW-5/VZ-1, Ansible BS-6. Only von-Vorteil items missing (VMware/Nutanix/Proxmox, 0 canonical). JD states militärische Karriere als höherer Unteroffizier oder Offizier von Vorteil, so BW-1 scores. Gated on the Frank Haugwitz call (comp, PSP eligibility, whether a second C5I application is welcome while 18027 is pending).",
-    "date": "2026-09-04"
-  },
   "https://jobs.ruag.ch/offene-stellen/senior-data-engineer-mssql-mit-data-lakehouse-perspektive/96da2356-d9f4-45da-a0ed-caa1acbe8ef9": {
     "company": "",
     "title": "",
@@ -2496,13 +2468,6 @@
     "note": "Same requisition JRQ$540-19848 he applied to and was rejected without interview (see applications.json); this URL is a repost UUID. NOTE 2026-10-08: posting requires 'Aktive Kenntnisse einer zweiten Amtssprache' - plausible, unproven factor in the rejection.",
     "date": "2026-10-08"
   },
-  "https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105": {
-    "company": "",
-    "title": "",
-    "decision": "maybe",
-    "note": "TRIAGE 2026-10-08 (JD read), NOT a user decision. TOP FIND. JR2027105, posted ~2026-10-06, Switzerland REMOTE or Zurich. Build/operate automation for DGX Cloud inference/agentic endpoints and control plane. Direct: Python, Ansible config mgmt + CloudFormation (JD says IaC, config mgmt OR GitOps), Linux, containers, AWS, SLOs (BS-3), on-call/incident (SW-2), production ML inference in 24/7 fab (BS-1). Adjacent: 8+ yrs production services and large-scale distributed systems; Kubernetes historical; metrics/logs/traces only via BS-4 PoC; networking fundamentals. Stand-out items (vLLM/SGLang/CUDA/Terraform/ArgoCD) are Gap but optional. Rough Evidence Fit ~68 provisional. CAVEAT: NVIDIA Senior DevOps JR2024256 pending since 2026-09-05 - stacking a second cold NVIDIA application is the user's call.",
-    "date": "2026-10-08"
-  },
   "https://www.google.com/about/careers/applications/jobs/results/110222556983632582-software-engineer-iii-site-reliability-engineering-prod-ai?location=Switzerland": {
     "company": "",
     "title": "",
@@ -2901,5 +2866,159 @@
     "decision": "skip",
     "note": "Claude triage 2026-10-22, JD read (reports/2026-10-22-fit-sources/nordic_ate_test_engineer.txt). Chip production test: 'Proven track record in high frequency, mixed-signal, digital and analog test development and execution' and 'Automated Test Equipment (ATE) Test Development in high volume manufacturing' - Gap. Bosch fab data analysis is adjacent domain, not ATE test development.",
     "date": "2026-10-22"
+  },
+  "https://jobs.apple.com/en-us/details/200634821-4170/aiml-senior-ml-rl-training-infrastructure-engineer-afm?team=MLAI": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/apple_afm_training_infra.txt). Apple Core Foundation Models, Zurich: RL training infra on TPU/JAX. Minimums 'Hands-on experience designing, building, or maintaining large-scale ML training infrastructure' and 'Strong proficiency with PyTorch or JAX and experience running training workloads on GPUs/TPUs' are Gap: BS-1 is inference integration, PyTorch is certification-context-only, JAX/TPU absent. Model-training infra at a top lab, outside the AI-platform overlap (user_positioning).",
+    "date": "2026-10-22"
+  },
+  "https://jobs.apple.com/en-us/details/200680814-4170/staff-machine-learning-engineer-siri-runtime-systems-and-interaction?team=MLAI": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, apple_siri_runtime_staff.txt). Despite the 'Runtime Systems' title the role leads audio/video GENERATION: minimum 'Deep hands-on experience with generative audio and/or video architectures (e.g., diffusion models, autoregressive models, GANs, VAEs...)' - Gap. Model-building.",
+    "date": "2026-10-22"
+  },
+  "https://jobs.apple.com/en-us/details/200687930-4170/hid-algorithms-software-engineer?team=HRDWR": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, apple_hid_algorithms.txt). HID sensor-algorithm frameworks with firmware teams. Minimum 'Strong C/C++ programming' - C++ is not a strength (feedback_cpp_emphasis); embedded preferred (ED-1 coursework only). SW architecture minimum would be Adjacent (BS-1/BS-2/SW-7), but the language gate decides.",
+    "date": "2026-10-22"
+  },
+  "https://jobs.apple.com/en-us/details/200634000-4170/applied-machine-learning-engineer-security?team=SFTWR": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, apple_applied_ml_security.txt). SEAR offensive security research. Minimums 'Expertise in ML, especially large language models and generative modeling' and offensive-security experience - both Gap (SW-8 is configured assistants; SW-5 Security Champion is a team role).",
+    "date": "2026-10-22"
+  },
+  "https://jobs.sbb.ch/v2/offene-stellen/projektleiter-in-datenmanagement/9eb0bb3e-89c7-4d3e-a892-8c6f587e0624": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, sbb_projektleiter_datenmanagement.txt). Job ID 104238, Olten. 'Diese Stelle befindet sich im Anforderungsniveau H' - below Niveau K, which the user ruled too low (2026-09-11). Project coordination of RIS/BIM inventory data capture and external suppliers, not engineering.",
+    "date": "2026-10-22"
+  },
+  "https://career2.successfactors.eu/sfcareer/jobreqcareer?jobId=1476&locale=de_DE&company=Swissgrid": {
+    "company": "Swissgrid (Aarau)",
+    "title": "(Senior) Cloud Security Engineer (80-100%)",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/swissgrid_cloud_security.txt); resolves the 07-06 title-only maybe. Azure cloud-security role in Platform Services, Aarau. Required 'Mehrjaehrige Erfahrung in Cyber Security Rollen mit Fokus auf Cloud Security', 'Erfahrung in Offensive Security (Security Auditing, Pentesting, CTF...)', Azure services and Terraform/GitHub Actions - Gap (Azure and Terraform forbidden; SW-5 Security Champion is a team role).",
+    "date": "2026-10-22"
+  },
+  "https://jobs.ruag.ch/offene-stellen/software-engineer-c5i/d4cff80d-97d5-4ef7-8460-fc72c54eb3be": {
+    "company": "RUAG (Thun/Bern)",
+    "title": "Software Engineer C5I m/f/d",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, ruag_software_engineer_c5i.txt); resolves the 07-06 title-only maybe. Thun, Fuehrungsinformationssysteme, non-senior. Required 'Sehr gute Kenntnisse von .NET Framework (C#), Visual Studio (ganzer Technologie Stack) sowie in SCRUM' - C#/.NET is limited/secondary (Fraunhofer only). Technical contact Cedric Wyss, Principal Software Engineering C5I.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/b17ef550-ab1f-44bf-b2c3-1adc9de1a54a": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_kredittbanken_data.txt). Kredittbanken Data Engineer, Trondheim. Requirement sentence 'God skriftlig og muntlig fremstillingsevne paa norsk og engelsk' = Norwegian required (A2); the scout read ENGLISH OK from the NAV 'Arbeidsspraak Engelsk eller norsk' field, a classifier defect. Stack is Azure Synapse/Power BI/DAX (Azure forbidden).",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/a1ce9d86-a95a-495e-9653-e1db8d67c100": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_kongsberg_norcontrol_devops.txt). Kongsberg Norcontrol DevOps & Test Engineer, Horten. Function fits (CI/CD, test automation, Docker/K8s), but requirement 'Kommunisere tydelig paa norsk og engelsk' = Norwegian required; scout read ENGLISH OK from the Arbeidsspraak field (classifier defect). Clearance required (not a gate). Kongsberg Seatex application pending.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/f9753f42-1af7-4282-9161-d3e52a8d1c62": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_kd_fullstack_positioning.txt). Kongsberg Discovery HiPOS full-stack, Horten; English only. Required 'application development using technologies such as C++, C#' and 'Real-time programming, real time operating systems' - Gap. Same company as the pending Seatex application; deadline 2026-10-23.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/9c0de8f3-3c34-47f7-aac3-b9cc359aea33": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_norwegian_fullstack.txt). Norwegian Air Shuttle Fullstack, Bodoe. 'Good communication skills, in Norwegian and English' = Norwegian required (scout read ENGLISH OK from the Arbeidsspraak field). Java/Kotlin/C# web and app development.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/cab34bc3-7ba6-48f6-b71c-5446a659a916": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_reitan_dataplattform.txt). Reitan Eiendom data engineer, Trondheim; function fits (DWH, ETL/ELT, Python, Databricks/Snowflake-type), but 'god muntlig og skriftlig fremstillingsevne paa norsk' = Norwegian required.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/3a90f1b5-2b87-4907-a82d-64411d58964a": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_norautron_dataplattform.txt). Norautron AI & Data Engineer, Horten: Microsoft Fabric/Azure/Power BI/Boomi platform - Azure forbidden, Fabric/Boomi absent. Language unstated (Norwegian ad).",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/7b71105c-1e7a-468b-aece-b03012651374": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_slb_software.txt). SLB Rig Equipment Software Development Engineer, Kristiansand; English only. Core stack 'C#, .NET, WPF, WinUI' desktop/equipment software - C#/.NET is secondary (Fraunhofer only). The sibling Data Scientist req is the better SLB fit.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/0edf3ca5-e911-4bbc-84e6-3f7ccb0c1131": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_kongsberg_maritime_senior_dev.txt). Kongsberg Maritime Drives R&D, Bergen: 'Strong understanding of drive control systems, power electronics, electric motors' - electrical/control engineering, not software.",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/ba3282b1-f394-4c95-85ac-38611d2f277e": {
+    "company": "",
+    "title": "",
+    "decision": "maybe",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_mazemap_senior.txt), NOT a user decision. MazeMap Senior Product Development Engineer, Trondheim; 'Fluent English Communication' only (Arbeidsspraak Engelsk). 8+ yrs IC building scalable backend - Direct/Adjacent (11 yrs; scale unquantified). 'Python Mastery: Deep, hands-on Python expertise (AsyncIO, performance tuning, microservices)' - Python Direct, AsyncIO specifics not in KB, ask. Java working knowledge (allowed-with-context), C++ exposure. AI dev tools - SW-8 (Copilot/Kiro). 'spatial data, graph algorithms, mathematical modeling' - BS-8 SDO_GEOMETRY + CRS transforms is a genuine bridge. ~21-person SaaS; reputation UNCHECKED (Glassdoor 1 review).",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/6c3365ed-adb8-4c64-8508-804d98df1037": {
+    "company": "",
+    "title": "",
+    "decision": "maybe",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_kda_platform.txt), NOT a user decision. KDA Defence Systems (NASAMS) platform developer, Kongsberg (travel Asker/Moss). Required: CI/CD and build/test/deploy automation (GN-1, FC-1, BS-6 Jenkins/GitLab), IaC and environment standardisation (BS-6 Ansible), Linux (BS-6, PP-3) - Direct. Preferred GitLab, containers (SW-3/BS-1), observability (BS-4 PoC only), CMake (unknown), embedded/Ada (ED-1 coursework). Clearance-eligible required (not a gate). LANGUAGE UNCLEAR: Norwegian ad, no requirement sentence, NAV tag norsk - ask the recruiter. Same group as the pending Kongsberg Seatex application (different business area) - stacking is the user's call. Kongsberg verdict OK (no Glassdoor; Indeed 3.8/101).",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/80658525-2197-4705-8d34-8f1b7535a5a8": {
+    "company": "",
+    "title": "",
+    "decision": "maybe",
+    "note": "TRIAGE 2026-10-22 (JD read, reports/2026-10-22-fit-sources/nav_sparebank1_ai.txt), NOT a user decision. SpareBank 1 Nord-Norge KI-ingenioer, Tromsoe. Required lines are generic dev (systems development, APIs/integrations, version control/testing, production lifecycle, security/privacy) - Direct. GenAI/agents/Azure/Copilot Studio are advantages only ('Erfaring med KI-utvikling er en fordel, men ikke et krav'); SW-8 forbids agent-orchestration claims. LANGUAGE UNCLEAR (Norwegian ad, no sentence, NAV tag norsk; a regional bank likely works in Norwegian). Tromsoe is far north. Reputation UNCHECKED.",
+    "date": "2026-10-22"
+  },
+  "https://jobs.ruag.ch/offene-stellen/senior-devops-kubernetes-engineer-c5i/4767c36f-2816-463f-846d-4f0283dc7c15": {
+    "company": "",
+    "title": "",
+    "decision": "parked",
+    "note": "USER DECISION 2026-10-22: not applying for now - 'there homeoffice rules are not good' (same C5I unit rejected 18027 after interview over home office). Keep it visible in the scout: 'maybe i need it later'. Fit unchanged (Sept estimate 86; JD re-fetched 2026-10-22, still live, no citizenship line).",
+    "date": "2026-10-22"
+  },
+  "https://arbeidsplassen.nav.no/stillinger/stilling/d2560ea2-eef6-45e7-babe-f94f9f236b6b": {
+    "company": "",
+    "title": "",
+    "decision": "skip",
+    "note": "USER DECISION 2026-10-22: 'SLB doesn't attract me at the moment'. Fit was real (data-engineering body, English only, Glassdoor 3.9); deadline 2026-10-29. Not a Norway-lane closure.",
+    "date": "2026-10-22"
+  },
+  "https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Zurich/Senior-Software-Developer_JR2008695": {
+    "company": "NVIDIA",
+    "title": "Senior Software Developer",
+    "decision": "skip",
+    "note": "Claude triage 2026-10-22, JD read (reports/2026-10-22-fit-sources/nvidia_senior_software_developer.txt): AI networking acceleration library; requires '8 years experience in modern C++ / C / Rust development' and 'Deep knowledge of the TCP/IP network stack' - both Gap (C++ is secondary per feedback_cpp_emphasis; networking 0 canonical).",
+    "date": "2026-10-22"
+  },
+  "https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105": {
+    "company": "",
+    "title": "",
+    "decision": "applied",
+    "note": "SUBMITTED 2026-10-22 via Workday. Evidence Fit 65 Stretch, hard gate PASS, Document Quality 85 (pre-edit; Edit 1 applied, not re-measured), channel Weak/cold, no cover letter. Package: output/NVIDIA_ProdEng_DGX_Cloud/.",
+    "date": "2026-10-22"
   }
 }

job_scout/state/employer_reputation.json

@@ -757,6 +757,55 @@
         "reviews": 18,
         "recommend_pct": 77
       }
+    },
+    "SLB": {
+      "aliases": [
+        "Schlumberger",
+        "SLB OneSubsea",
+        "OneSubsea"
+      ],
+      "glassdoor": {
+        "rating": 3.9,
+        "reviews": 15113,
+        "recommend_pct": 73,
+        "scope": "company-wide"
+      },
+      "note": "Work-life balance 3.4 (lowest sub-rating), culture 3.9, career 3.8; praise for training and international mobility, criticism of long hours and management. Industry average 3.7.",
+      "sources": [
+        "https://static.glassdoor.com.br/Reviews/SLB-Reviews-E588_P3871.htm"
+      ],
+      "checked": "2026-10-22"
+    },
+    "MazeMap": {
+      "aliases": [
+        "Mazemap As",
+        "MazeMap AS"
+      ],
+      "glassdoor": {
+        "rating": null,
+        "reviews": null
+      },
+      "verdict_override": "unchecked",
+      "note": "Glassdoor page exists with a single review (3.0, March 2024: good colleagues and office; no full remote, sales over engineering). One review is no signal, so the verdict stays unchecked. ~21 employees (OpenCorporates). Do not confuse with 'Maze' (E4132839).",
+      "sources": [
+        "https://www.glassdoor.com/Reviews/MazeMap-Reviews-E4814325.htm",
+        "https://opencorporates.com/companies/no/912197905"
+      ],
+      "checked": "2026-10-22"
+    },
+    "SpareBank 1 Nord-Norge": {
+      "aliases": [
+        "SpareBank 1 Nord Norge",
+        "Sparebanken Nord-Norge"
+      ],
+      "glassdoor": {
+        "rating": null,
+        "reviews": null
+      },
+      "verdict_override": "unchecked",
+      "note": "No Glassdoor, Great Place to Work or Universum entry found 2026-10-22. Bytt.no's 2.5 is a CUSTOMER rating, not an employee one, and must not be used.",
+      "sources": [],
+      "checked": "2026-10-22"
     }
   }
 }

knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md

@@ -0,0 +1,66 @@
+# On-duty alerting, networking and Bosch GPUs — user clarification
+
+Source: direct user message, 2026-10-22, during the NVIDIA *Senior Production Engineer, DGX Cloud*
+(JR2027105) fit check. Three questions were asked because the KB has repeatedly under-recorded
+ordinary operations work (Ansible, Linux, Bash, Kotlin, Jenkins/GitLab).
+
+## Questions and verbatim answers
+
+1. *Monitoring in your current Swisscom unit: do you monitor or get alerted on your Glue/Airflow
+   pipelines (CloudWatch alarms, failure alerts, dashboards)? Is there on-call in your current team?*
+
+   > Yes, we get alerted ... we are using opsgenie and we have a engineer on duty role which rotates.
+   > Bugs and Alerts also automatically create Jira issue tickets which we dann asign an fix.
+
+2. *Networking at work: any AWS VPC, security group or private-endpoint work at Swisscom, or firewall,
+   DNS, proxy or load-balancer work at Bosch?*
+
+   > No, not really - probably only my AWS solution architect cert covers AWS network topics... but day
+   > to day work is using automated landing zones.
+
+3. *GPUs at Bosch: did the ML inference containers run on GPUs?*
+
+   > The ML inference did ran on NVIDIA GPUs (using CUDA for performance boost - but i never did much
+   > configuration there)
+
+## Follow-up questions and verbatim answers (same day)
+
+4. *Bosch monitoring PoC: your earlier CV says it also included Grafana, Prometheus and Loki; the Bosch
+   reference letter only mentions ELK. Correct?* and 5. *Do you also write or adjust the alerts for your
+   own pipelines, or only respond to them?*
+
+   > 1. both are correct 2. yes, i wrote e.g. thresholds for data quality etc....
+
+## Established facts
+
+- **BS-4 stack:** the Bosch PoC covered ELK and Kafka plus monitoring and alerting with Grafana,
+  Prometheus and Loki. It is still a proof of concept.
+- **SW-11 alert definitions:** he defines alert thresholds for his own pipelines; data-quality
+  thresholds are the named example. Other alert types are not established.
+
+- **Current on-duty (new claim SW-11):** his current Swisscom team uses Opsgenie for alerting and has a
+  rotating engineer-on-duty role that he takes part in. Bugs and alerts automatically create Jira
+  tickets, which the team assigns and fixes. This is the current counterpart of the past SW-2 on-call
+  at ART B2B.
+- **Networking (skill `networking`, certification context only):** no professional network
+  configuration. AWS networking is covered by the AWS Certified Solutions Architect – Associate
+  (26 Sep 2024, active). At work he uses automated AWS landing zones that the platform provides.
+- **Bosch GPUs (BS-1 scope):** the containerized ML inference he integrated ran on NVIDIA GPUs, with
+  CUDA used for acceleration. He did little CUDA or GPU configuration himself.
+
+## Not established
+
+- Who built or configured the alerting, the Opsgenie routing or the Jira automation.
+- Which metrics or logs feed the alerts; SLIs, SLOs or error budgets at Swisscom.
+- Incident counts, response times, MTTR or availability figures.
+- Whether the on-duty role covers nights and weekends (do not write 24/7).
+- Any professional VPC, security-group, firewall, DNS, proxy or load-balancer work.
+- Building or designing landing zones or network topology.
+- CUDA programming or configuration, GPU tuning or GPU performance analysis. GPU model, count and
+  serving framework at Bosch.
+
+## Canonical destination
+
+`claims.json`: new claim **SW-11**; **BS-1** scope and forbidden list; skills **Opsgenie**
+(allowed-with-context), **CUDA** (forbidden; system context only) and **networking**
+(certification-context-only).

output/NVIDIA_ProdEng_DGX_Cloud/JD_nvidia_prod_eng_dgx_cloud_jr2027105.txt

@@ -0,0 +1,97 @@
+Source: https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105
+Retrieved: 2026-10-22; HTTP 200; Playwright rendered
+
+Skip to main content
+CAREERS AT NVIDIA
+Sign In
+Home
+Search for Jobs
+Senior Production Engineer - DGX Cloud page is loaded
+Senior Production Engineer - DGX Cloud
+Apply
+locations
+Switzerland, Remote
+Switzerland, Zurich
+time type
+Full time
+posted on
+Posted 16 Days Ago
+job requisition id
+JR2027105
+
+NVIDIA DGX Cloud delivers AI services and endpoints for research and production workloads. We are looking for a Senior Production Engineer to build software and automation that make those services reliable, scalable, and safe to operate. The Production Engineering team works on large-scale distributed systems spanning internal and external model endpoints; regional control plane services that orchestrate workloads and route requests; and the GPU/CPU compute infrastructure where inference and agentic workloads run. Our work spans Kubernetes clusters across AWS, Azure, Google Cloud, other partner cloud environments, and on-premises deployments.
+
+
+
+
+What you’ll be doing:
+
+Build and operate production software, automation, and tooling for control plane services, model deployments, and inference and agentic workloads across DGX Cloud environments.
+Improve the reliability of inference and agentic platforms and services, including NVIDIA Cloud Functions, SGLang- and vLLM-based endpoints, and inference services built with NVIDIA Dynamo, through health validation, safer rollouts, observability, and recovery.
+Improve endpoint availability, inference routing, capacity management, and service health to maintain predictable performance as workloads and demand change.
+Use infrastructure as code and GitOps to deploy, configure, validate, upgrade, and recover services consistently across environments.
+Build workflows for service enablement, model releases, handoff, deprecation, and ongoing operations; replace repeatable manual work with reliable automation.
+Define and instrument SLIs and SLOs for inference and control plane services, including availability and latency, use error budgets to guide reliability improvements, and make production health visible to partner teams.
+Participate in on-call and incident response, troubleshoot failures across routing, model runtimes, software, and infrastructure, and turn recurring issues into automation and durable fixes.
+Collaborate with model, platform, storage, networking, security, and GPU infrastructure teams to design and operate services safely at scale.
+
+
+
+What we need to see:
+
+8+ years of experience building or operating production services and large-scale distributed systems, including hands-on automation.
+Strong programming skills in Python, Go, or a comparable language, with experience developing tools for production operations.
+Experience with infrastructure as code, configuration management, or GitOps, and with building automation for repeatable service deployments and changes.
+Strong knowledge of Linux, Kubernetes, containers, cloud infrastructure, distributed systems, and networking fundamentals; ability to diagnose failures in production.
+Understanding of Production Engineering principles, including SLIs, SLOs, error budgets, incident response, and reducing operational toil.
+Experience instrumenting services and using metrics, logs, and traces to understand system behavior and improve reliability.
+Clear technical communication and ability to work across engineering teams.
+BS/MS in Computer Science or equivalent practical experience.
+
+
+
+Ways to stand out from the crowd:
+
+Familiarity with technologies such as vLLM, SGLang, PyTorch, TensorRT-LLM, NVIDIA Dynamo, CUDA, or NCCL, and with GPU performance analysis.
+Experience building Kubernetes operators, controllers, workload orchestration services, fleet management systems, or self-healing automation.
+Experience with Terraform, Argo CD, CI/CD, policy validation, or safe deployment and rollback systems.
+Background with developing with AI tools and agents.
+Experience with production AI inference or agentic workloads, including debugging issues across models, runtimes, Kubernetes, and hardware.
+
+
+
+NVIDIA is leading the way in groundbreaking developments in Artificial Intelligence, High-Performance Computing and Visualization. The GPU, our invention, serves as the visual cortex of modern computers and is at the heart of our products and services. We have some of the most forward-thinking and hard-working people on the planet working for us. If you're creative, hard-working and self-motivated, we want to hear from you!
+
+Similar Jobs (4)
+Senior Software Engineer, JAX
+locations
+2 Locations
+time type
+Full time
+posted on
+Posted 3 Days Ago
+Senior AI Compute Engineer
+locations
+4 Locations
+time type
+Full time
+posted on
+Posted 30+ Days Ago
+Senior Software Architect, GPU Networking Research
+locations
+2 Locations
+time type
+Full time
+posted on
+Posted 30+ Days Ago
+View All 4 Jobs
+
+NVIDIA pioneered accelerated computing. Today, our AI infrastructure powers global intelligence, transforming every industry.
+
+ 
+
+Learn more about NVIDIA.
+
+Follow Us
+Applicant Privacy Policy
+© 2026 Workday, Inc. All rights reserved.

output/NVIDIA_ProdEng_DGX_Cloud/critique_nvidia_prod_eng_dgx_cloud.md

@@ -0,0 +1,96 @@
+# Critique: NVIDIA — Senior Production Engineer, DGX Cloud (JR2027105)
+
+> **STALE as of Edit 1 (2026-10-22):** Tier 1 #1 and Tier 2 #1–#4 were applied (see the session's Edit History). The Document Quality score of 85 measures the pre-edit document and was not re-measured.
+
+Date: 2026-10-22. Documents: `e2e_nvidia_prod_eng_dgx_cloud_resume.tex` / `.pdf` (2 pages). No cover letter (session decision NO).
+JD: `JD_nvidia_prod_eng_dgx_cloud_jr2027105.txt`, verbatim Playwright render 2026-10-22, live.
+
+## 1. Fit verdict and hard gates
+
+- **Fit class: Stretch.** **Hard gate: PASS.** No required line is a Gap; location (Switzerland Remote), language and authorization are resolved.
+- Stretch because required Q5 (SLIs, SLOs, error budgets, incident response, toil), the SRE core of the title, is only Adjacent: SLOs (BS-3) and incident response (SW-2 past, SW-11 current) are evidenced; SLIs and error budgets are not.
+- Likeliest screen rejection: Q1 *"8+ years of experience building or operating production services and large-scale distributed systems"*. The years hold (visible record from 2015-05); the large-scale distributed systems do not at NVIDIA's scale.
+
+## 2. Evidence Fit — 65/100 (scored before reading the prose; unchanged from the session)
+
+| Dimension | Weight | Score | Basis |
+|---|---:|---:|---|
+| Required qualifications | 35 | 26 | 4 Direct (Q2 Python/tools, Q3 IaC/config mgmt, Q7 communication, Q8 degree); 4 Adjacent (Q1 scale, Q4 networking, Q5 SLIs/error budgets, Q6 instrumenting beyond alert thresholds + PoC) |
+| Core responsibilities | 25 | 12 | On-call/incident Direct and now current (SW-11); inference integration Adjacent (BS-1); routing/capacity Gap |
+| Level and ownership | 15 | 11 | Senior IC vs Staff title; component and application ownership |
+| Recency and depth | 10 | 6 | Current: on-duty, alert thresholds, AWS/CloudFormation, Python. Past: Ansible/Linux (2022), Kubernetes, GitLab |
+| Domain/tool transfer | 10 | 5 | Production inference on NVIDIA GPUs is a real bridge; vLLM/SGLang/Dynamo, K8s operators, Terraform, Argo CD do not transfer |
+| Practical constraints | 5 | 5 | Remote CH, English, EU/B permit |
+| **Total** | | **65** | |
+
+## 3. Document Quality — 85/100
+
+| Dimension | Weight | Score | Note |
+|---|---:|---:|---|
+| Truth and provenance | 25 | 23 | Every claim traces to canonical scope; one over-generalisation ("set **the** alert thresholds", Tier 1 #1) |
+| Information hierarchy | 20 | 17 | Clear headers; NVIDIA-GPU bullet on page 1; experience starts ~44% down page 1. The Swisscom tense switch (present → past for ART B2B work) is unexplained to a reader |
+| Bullet evidence and impact | 20 | 15 | Specific and scoped, no invented metrics (none exist). Two dense bullets: Swisscom #1 (38 words, "alerts" twice) and Bosch #1 ("GPU-backed" and "on NVIDIA GPUs" in one sentence) |
+| Relevance and terminology | 15 | 12 | Honest coverage of IaC, configuration management, Linux, Kubernetes, containers, AWS, SLOs, incident response, on-call, alerting, CI/CD, inference, GPU. Absent: "observability" (honestly available for the BS-4 PoC); SLIs, error budgets, networking, Terraform, GitOps, CUDA correctly absent |
+| Skills evidence | 10 | 9 | All canonical; Kubernetes is listed without a recency cue (bullets show it as past) |
+| Mechanics/readability | 10 | 9 | Validator PASS, 0 cadence warnings, 0 boxes, 2 pages, parse order correct |
+
+## 4. Channel Strength — Weak
+
+Cold portal application; no contact published. The warm action (LinkedIn note to an NVIDIA recruiter or a DGX Cloud Production Engineering engineer in Switzerland) is the cheapest improvement available and is not yet done.
+
+## 5. Competitive read and reader sequence
+
+- **Obvious-fit competitor:** an SRE/production engineer from a hyperscaler, GPU cloud or ML-serving company with vLLM/Triton-style endpoints on Kubernetes, operators, Terraform/Argo CD and formal SLO/error-budget practice.
+- **Dennis's real advantage:** ML inference on NVIDIA GPUs integrated into a fab where a defect could have stopped production; Ansible depth including his own extensions; on-call across two Swisscom units, current with alert thresholds he writes; current AWS IaC; ten years of automation across five employers.
+- **Competitor's real advantage:** inference-serving stack, Kubernetes at scale and current, SRE formalism, networking. Do not value the GPU bridge as highly as their direct serving experience.
+- **Parser/ATS:** extracts cleanly; strong on IaC/config mgmt/Linux/K8s/AWS/SLO/on-call/alert/GPU literals; no "distributed systems" phrase beyond "distributed processing" and "distributed video-transcoding".
+- **Recruiter glance: Borderline.** Data-engineer identity in the title line against a production-engineering req; the on-duty bullet and the GPU line are what carry it past a skim.
+- **Hiring manager: Maybe.** Real production operations and automation, real GPU inference in a harsh environment; no serving stack, Kubernetes not current.
+- **Technical reviewer — likely first challenges:** (1) "Walk me through a production incident you debugged end to end" (SW-2/SW-11); (2) "How did you roll out a new model version safely in the fab?" (BS-1 — the KB records the failure-consequence constraint, not the rollout mechanics; prepare the honest answer); (3) "When did you last run Kubernetes, and how would you debug a failing pod?"; (4) distributed systems at scale; (5) vLLM/SGLang/Dynamo — say plainly you have not used them.
+
+## 6. Claim audit
+
+| Claim | Canonical ID/evidence | Direct/Adjacent | Safe? | Fix |
+|---|---|---|---|---|
+| "over ten years in production software and its automation" | Employment from Generali 2015-05; Capgemini excluded | Direct | Yes | — |
+| Integrated containerized ML inference on NVIDIA GPUs into a continuously operating fab (summary + Bosch #1) | BS-1 incl. 2026-10-22 hardware note | Direct | Yes; no CUDA/GPU-tuning claim | Tier 2 #3 (redundancy) |
+| Builds AWS data pipelines with CloudFormation | SW-1, SW-7; CloudFormation production-current | Direct | Yes | Be ready to describe templates he wrote |
+| Serves in the engineer-on-duty rotation; Opsgenie; auto-created Jira tickets | SW-11 | Direct | Yes | — |
+| "set **the** alert thresholds, including data-quality checks" | SW-11 ("i wrote e.g. thresholds for data quality") | Direct | **Over-generalised** | Tier 1 #1 |
+| Migrated Fulfillment and Product Analysis domains' ETL to AWS as Python Glue jobs, Airflow, CloudFormation, contributing to wider migration | SW-1 (scoped object, hedged contribution) | Direct | Yes | — |
+| Design and build governed data products within the company-wide Data Mesh, unit tests | SW-7, SW-6, SW-9 | Direct | Yes; mesh not claimed | — |
+| Built and operated Python applications on Kubernetes via automated GitLab CI/CD, multiple environments | SW-3 (past; worked with CI/CD, did not build it) | Direct (past) | Yes | Tier 2 #1 (tense cue) |
+| Owned Fulfillment ETL components, on-call, 2nd/3rd-level support, RCA | SW-2 (past tense) | Direct (past) | Yes | — |
+| Administered and automated Linux servers; Git-versioned Ansible run from Jenkins/GitLab | BS-6 (no GitOps) | Direct (past) | Yes | — |
+| Ansible credential plugin caching keystore secrets | BS-6 | Direct | Yes; unquantified | — |
+| Technical Application Owner for KLA DMS and Spotfire; defined SLOs | BS-3, BS-5 | Direct | Yes | — |
+| ELK/Kafka PoC with Prometheus, Grafana, Loki | BS-4 incl. 2026-10-22 confirmation | Adjacent (PoC) | Yes; PoC label kept | Tier 2 #4 (observability literal) |
+| Jenkins build jobs for automatic deployment (Fraunhofer) | FC-1 | Direct | Yes | — |
+| Test suites as CI/CD quality gates; contributed Python to distributed transcoding backend | VZ-2, VZ-1 (contributor; no customer names) | Direct | Yes | — |
+| XLDeploy migration; BDD test automation responsibility, Jenkins | GN-3, GN-1 | Direct | Yes | — |
+| Skills: PyTorch/TensorFlow inside certificate line | certification-context-only | — | Yes (validator) | — |
+| Contact: EU citizen, Swiss B permit | identity | — | Yes | — |
+
+## 7. Fixes
+
+### Tier 1 — truth/fit
+1. **Swisscom #1:** "set **the** alert thresholds" → "set alert thresholds". He named data-quality thresholds as an example; "the" implies he defines all of them.
+
+### Tier 2 — presentation
+1. **Swisscom tense switch:** prefix the two past bullets (Kubernetes/GitLab; Fulfillment ETL on-call) with "Earlier in the role," so the present → past change reads as intentional. No dates or unit names (never date the ART units).
+2. **Add a one-line personal-infrastructure entry (PP-3):** "Personal: self-host and harden a Debian server (~10 years) running nginx, a mail server, a VPN, Docker services and Bitwarden." It is the only current Linux evidence and the only networking-fundamentals evidence (reverse proxy, VPN, mail). Page 2 has room. PP-3 rules: personal label, no uptime/scale figure, never production/enterprise, never datacenter.
+3. **Tighten the two dense bullets:** Swisscom #1 (38 words, "alerts" twice) and Bosch #1 ("GPU-backed … deployed on NVIDIA GPUs" says GPU twice; keep "on NVIDIA GPUs").
+4. **Optional:** add the word "observability" to the BS-4 bullet ("observability proof of concept: …"), keeping the PoC label. It is the JD's own term and honest for that PoC.
+
+## 8. Cover letter
+
+Session decision NO: Workday portal, no letter requested, nothing the resume cannot show. Absence not penalised. The warm LinkedIn note has more information value than a letter.
+
+## 9. Mechanical verification
+
+- `validate_resume_system.py --document`: **PASS**, 0 warnings (cadence included).
+- MiKTeX pdflatex: **2 pages**, 0 overfull/underfull boxes. Both pages inspected at 80 dpi: no clipping or overlap; the BS-1 GPU bullet sits on page 1; page 2 ends about half-way (no filler added, by rule).
+- `pdftotext -layout`: header → Summary → Skills → Experience (Swisscom, Bosch, Fraunhofer, Vizrt, Generali, each employer/date then title/location) → Education. No Capgemini. No Go, Terraform, GitOps, CUDA or networking claim in the text.
+- Not run: `page_metrics.py` (its `pdftotext -bbox` call fails with the Git-for-Windows binary; page-1 experience start estimated visually at ~44%).
+
+**Not submit-ready until Tier 1 #1 is applied.** After that the package is a sound Stretch application, still Weak-channel.

output/NVIDIA_ProdEng_DGX_Cloud/e2e_nvidia_prod_eng_dgx_cloud_resume.tex

@@ -0,0 +1,82 @@
+% NVIDIA -- Senior Production Engineer, DGX Cloud (JR2027105), Switzerland Remote / Zurich.
+% International Tech profile. Written 2026-10-22 from resume_builder/canonical/claims.json
+% (incl. SW-11 and the BS-1/BS-4 additions of 2026-10-22); no historical output used as a source.
+
+\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 \& AI Engineer $\vert$ Production Automation and Operations $\vert$ AWS $\cdot$ ML Inference}
+\contactline{Bern, Switzerland $\vert$ EU citizen, Swiss B permit $\vert$ \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}
+Engineer with over ten years in production software and its automation. At Bosch, integrated containerized ML inference on NVIDIA GPUs into a continuously operating semiconductor fab; at Swisscom, builds AWS data pipelines with CloudFormation and serves in the engineer-on-duty rotation.
+\end{rSection}
+
+\begin{rSection}{Technical Skills}
+\skillline{Languages}{Python, SQL, Bash}
+\skillline{Infrastructure}{AWS, infrastructure as code (CloudFormation), configuration management (Ansible), Linux administration, Docker, Kubernetes, Jenkins, GitLab CI/CD}
+\skillline{Operations and data}{On-duty incident response, alert thresholds, SLOs, root-cause analysis, unit testing (pytest); PySpark (distributed processing), Kafka, Apache Iceberg}
+\skillline{AI tooling}{AI-assisted development with Kiro and Copilot}
+\skillline{Certifications}{AWS Certified Solutions Architect -- Associate (2024); IBM AI Engineering Professional Certificate, incl.\ PyTorch and TensorFlow courses (2020)}
+\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 Serve in the team's rotating engineer-on-duty role, responding to Opsgenie alerts and resolving issues through auto-created Jira tickets; set alert thresholds, including data-quality checks, for owned pipelines.
+\item Migrated the Fulfillment and Product Analysis domains' legacy ETL pipelines to AWS as Python Glue jobs orchestrated with Airflow and provisioned with CloudFormation, contributing to the wider company migration.
+\item Design and build governed data products on AWS (PySpark, Glue, Athena, Apache Iceberg) within Swisscom's company-wide Data Mesh, with unit tests for the pipeline code.
+\item Earlier in the role, owned the Fulfillment ETL components that moved business-critical data from Oracle into the Teradata data warehouse, including on-call duty, 2nd/3rd-level support and root-cause analysis.
+\item Built and operated Python applications on Kubernetes in the same period, deployed through automated GitLab CI/CD pipelines across multiple environments.
+\end{rSubsection}
+
+\begin{rSubsection}{Robert Bosch Semiconductor Manufacturing Dresden GmbH}{Feb 2020 -- Dec 2022}{Senior Data Engineer, Data Analysis (promoted from Data Engineer, Jan 2021)}{Dresden, Germany}
+\item Designed and executed the integration of containerized ML inference on NVIDIA GPUs (Docker, Kubernetes, Ansible) into a continuously operating semiconductor fab, where a software defect could have halted production; reduced manual image-based defect classification.
+\item Administered and automated the Linux servers behind fab analytics and ML workloads with Git-versioned Ansible playbooks run from Jenkins and GitLab pipelines.
+\item Developed Ansible extensions, including a credential plugin that cached keystore secrets locally to cut lookup volume.
+\item Served as technical Application Owner for the KLA Defect Management System and TIBCO Spotfire, defining SLOs and handling vendor communication, user training and documentation.
+\item Built an observability proof of concept for log management and anomaly detection on ELK and Kafka, with monitoring and alerting in Prometheus, Grafana and Loki.
+\end{rSubsection}
+
+\begin{rSubsection}{Fraunhofer Center for Maritime Logistics and Services CML}{Sep 2018 -- Oct 2019}{Research Software Engineer}{Hamburg, Germany}
+\item Set up Jenkins, including the build jobs for automatic deployment of a C\# and SQL Server crew-scheduling application.
+\end{rSubsection}
+
+\begin{rSubsection}{Vizrt}{Jul 2017 -- May 2018}{Test Automation / DevOps Engineer}{Bergen, Norway}
+\item Developed automated test suites wired into the CI/CD pipeline as quality gates, and contributed Python to a distributed video-transcoding backend.
+\end{rSubsection}
+
+\begin{rSubsection}{Generali Deutschland Informatik Services GmbH}{May 2015 -- Jun 2017}{IT Consultant}{Hamburg, Germany}
+\item Migrated web services to a new XLDeploy deployment procedure and held technical responsibility for BDD test automation (Serenity-BDD, JBehave), with its build jobs administered in Jenkins.
+\end{rSubsection}
+
+\end{rSection}
+
+\begin{rSection}{Personal Infrastructure}
+Self-host and harden a personal Debian server, run for about ten years: nginx, a mail server, a VPN server, Docker services and Bitwarden.
+\end{rSection}
+
+\begin{rSection}{Education}
+\compactentry{M.Eng. Computer Aided Engineering}{Apr 2012 -- Oct 2013}
+Universität der Bundeswehr München; Software Design \& Engineering.
+
+\compactentry{B.Eng. Information and Telecommunication Technologies}{Oct 2009 -- Oct 2012}
+Universität der Bundeswehr München.
+\end{rSection}
+
+\end{document}

output/NVIDIA_ProdEng_DGX_Cloud/resume.cls

@@ -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/NVIDIA_ProdEng_DGX_Cloud/session_nvidia_prod_eng_dgx_cloud.md

@@ -0,0 +1,162 @@
+# Session: NVIDIA — Senior Production Engineer, DGX Cloud (Switzerland Remote / Zurich)
+
+## JD Integrity
+- **File/source:** `output/NVIDIA_ProdEng_DGX_Cloud/JD_nvidia_prod_eng_dgx_cloud_jr2027105.txt` — req **JR2027105**
+  (`https://nvidia.wd5.myworkdayjobs.com/en-US/NVIDIAExternalCareerSite/job/Switzerland-Remote/Senior-Production-Engineer----DGX-Cloud_JR2027105`)
+- **Retrieval method and date:** headless Playwright render via `job_scout/.venv` (Workday SPA), **2026-10-22**, HTTP 200.
+  Same text first read in the 2026-10-08 scout triage.
+- **Verbatim posting:** YES — saved unedited.
+- **Posting status:** LIVE. *"Posted 16 Days Ago"* on 2026-10-22, so posted ~2026-10-06. Locations: **Switzerland, Remote** and Switzerland, Zurich.
+
+## Application Decision
+- **Audience profile:** International Tech (US-tech, English).
+- **Evidence Fit: 63/100** (dimension table below). The 2026-10-08 triage said "~68 provisional"; this is the full scoring and it is lower.
+- **Fit class: Stretch.** The score sits in the Adjacent band, but `application_strategy.md` makes a role Stretch when a defining
+  responsibility is only Adjacent, and `config.md`'s decision tree marks SRE as Stretch when it is a required qualification.
+  Required line Q5 (SLIs, SLOs, error budgets, incident response, toil) is the SRE core of a Production Engineer title, and only
+  part of it is evidenced. Judgment recorded: the function itself (operating and automating production services: SW-2, BS-1, BS-6)
+  is Direct, so this is Stretch, not a title-defining Gap / No-go. A strict reading of `critique_framework.md` §1 ("title-defining
+  capability is not Direct") would flag it harder.
+- **Hard gate: PASS.** No required line is a Gap. Two required lines are only Adjacent and carry the screening risk:
+  Q1 ("8+ years ... large-scale distributed systems") — **the likeliest recruiter-screen rejection** — and Q4 (networking fundamentals).
+- **Practical constraints: resolved.** Remote from Switzerland is stated in the posting (user baseline is 3 home-office days; fully remote
+  is acceptable when comp and role are strong). Comp: levels.fyi Greater Zurich NVIDIA IC3 median ~CHF 204k (recorded 2026-09-04 in
+  `output/NVIDIA_SrDevOps_Zurich/session_nvidia_sr_devops.md`; self-reported, equity-heavy). English only. EU citizen, B permit.
+- **Stacking:** NVIDIA Senior DevOps JR2024256 (applied 2026-09-05) was last listed in the 2026-09-18 scrape, absent since 2026-09-23,
+  page "doesn't exist" on 2026-10-22, no response. The req is pulled, so no open NVIDIA posting to stack against (recorded in the tracker).
+- **Channel Strength: Weak** (cold portal application).
+- **Channel plan:** no contact is published. Warm action: a short LinkedIn note to an NVIDIA recruiter or a DGX Cloud Production Engineering
+  engineer in Switzerland/Zurich, before or right after submitting.
+- **Named contact published?** NO.
+- **Decision: PROCEED (Stretch), 2026-10-22.** Fit gate presented; the user answered the evidence questions and confirmations and kept the instruction to prepare the CV. Recorded as Stretch, not upgraded by the go-ahead.
+
+### Evidence Fit — dimension table
+| Dimension | Weight | Score | Basis |
+|---|---:|---:|---|
+| Required qualifications | 35 | 26 | 4 Direct (Q2, Q3, Q7, Q8), 4 Adjacent (Q1, Q4, Q5, Q6), 0 Gap |
+| Core responsibilities | 25 | 12 | 1 Direct (on-call/incident, SW-2), 6 Adjacent, 1 Gap (inference routing/capacity) |
+| Level and ownership | 15 | 11 | Senior IC vs current Staff title; component/application ownership (SW-2, BS-3); distributed-systems scale is the weak part |
+| Recency and depth | 10 | 5 | The ops core is past: Ansible/Linux at Bosch to 2022; Kubernetes, GitLab CI/CD and on-call at ART B2B. Current: AWS, CloudFormation, Python/PySpark, Docker, personal Linux (PP-3) |
+| Domain/tool transfer | 10 | 4 | IaC and CI/CD transfer; vLLM/SGLang/Dynamo/NVCF, GPU, Kubernetes operators, Terraform, Argo CD do not |
+| Practical constraints | 5 | 5 | Remote CH, English, authorization |
+| **Total** | 100 | **63** | |
+
+## Requirements
+
+### Required — "What we need to see"
+| # | Requirement | Class | Canonical evidence (currency) | Gate? |
+|---|---|---|---|---|
+| Q1 | 8+ years building or operating production services and large-scale distributed systems, including hands-on automation | **Adjacent** | Years: ~9.8 years of visible employment (Generali 2015-05 onward, Capgemini excluded), each with production software plus automation: GN-1, VZ-1/VZ-2, FC-1, BS-1/BS-6, SW-1/SW-2/SW-3. "Large-scale distributed systems": SW-1/SW-6 Glue/PySpark/Athena on AWS (current, managed distributed processing); SW-3 Kubernetes and Kafka (past); VZ-1 contributor to a distributed transcoding backend. Not NVIDIA-scale | **Screen risk** |
+| Q2 | Strong programming in Python, Go or comparable; tools for production operations | **Direct** | Python production-current; BS-6 Ansible extensions incl. credential-caching plugin (past); SW-1 migration automation as Python Glue jobs (current) | — |
+| Q3 | IaC, configuration management, or GitOps; automation for repeatable service deployments and changes | **Direct** | Ansible BS-6 (past, intensive); CloudFormation production-current; FC-1 Jenkins build jobs for automatic deployment; SW-3 deployed through automated GitLab CI/CD across multiple environments (past, worked with, did not build). **GitOps forbidden (BS-6)** — the "or" is met without it | — |
+| Q4 | Strong knowledge of Linux, Kubernetes, containers, cloud, distributed systems, networking fundamentals; diagnose production failures | **Adjacent** | Linux BS-6 (past) + PP-3 (current, personal); Kubernetes SW-3/BS-1 (past only); Docker current; AWS current; distributed systems as Q1; **networking: PP-3 personal only (nginx, VPN, mail) — open question**; diagnosis SW-2 root-cause analysis, 2nd/3rd-level support (past) | Watch |
+| Q5 | Production Engineering principles: SLIs, SLOs, error budgets, incident response, reducing toil | **Adjacent** | SLOs BS-3 (past); on-call and incident handling SW-2 (past); toil reduction through automation BS-6, SW-1, SW-3. **SLIs and error budgets not established.** BS-6 forbids an SRE role or title claim | Defines Stretch |
+| Q6 | Instrumenting services; metrics, logs, traces | **Adjacent** | BS-4 ELK/Kafka PoC plus monitoring (past, PoC — never a platform). **Current monitoring at Swisscom unknown — open question** | — |
+| Q7 | Clear technical communication; work across engineering teams | **Direct** | SW-4 stakeholders/architects/engineers; BS-3 vendors and internal customers; BS-5 TAF 2022 co-presentation; GN-1 presenting BDD; BS-9/SW-10 onboarding juniors | — |
+| Q8 | BS/MS in Computer Science or equivalent | **Direct** | EDU-MENG (Software Design & Engineering focus), EDU-BENG | — |
+
+### Core responsibilities — "What you'll be doing"
+| # | Responsibility | Class | Evidence |
+|---|---|---|---|
+| R1 | Build/operate production software, automation, tooling for control plane, model deployments, inference and agentic workloads | Adjacent | BS-1 designed and executed containerized ML inference integration in a 24/7 fab (past). Control plane and agentic: none (SW-8 forbids agent orchestration) |
+| R2 | Reliability of inference platforms (NVCF, SGLang/vLLM, Dynamo): health validation, safer rollouts, observability, recovery | Adjacent | BS-1 failure-consequence constraint (a defect could have stopped production) is real safe-rollout motivation; serving stacks are Gap |
+| R3 | Endpoint availability, inference routing, capacity management | Gap | — |
+| R4 | IaC and GitOps to deploy/configure/validate/upgrade/recover across environments | Adjacent | IaC Direct (Q3); GitOps never |
+| R5 | Workflows for service enablement, model releases, handoff, deprecation; automate manual work | Adjacent | SW-1 migration automation; SW-7 onboarding sources and data products; BS-1 |
+| R6 | Define/instrument SLIs and SLOs, error budgets, visible production health | Adjacent | BS-3 SLOs |
+| R7 | On-call, incident response, troubleshooting; recurring issues into automation | **Direct** | SW-2 (past) |
+| R8 | Collaborate with model, platform, storage, networking, security, GPU teams | Adjacent | Cross-team work Direct (Q7); these specific teams not |
+
+### Preferred — "Ways to stand out"
+| # | Item | Class |
+|---|---|---|
+| P1 | vLLM, SGLang, PyTorch, TensorRT-LLM, Dynamo, CUDA, NCCL; GPU performance analysis | Gap (PyTorch is certification-context only) |
+| P2 | Kubernetes operators/controllers, workload orchestration, fleet management, self-healing automation | Gap (BS-6 Linux-estate automation is not fleet management at this level) |
+| P3 | Terraform, Argo CD, CI/CD, policy validation, safe deployment and rollback | Partial: CI/CD Direct (Jenkins FC-1/GN-1, GitLab SW-3/BS-6), VZ-2 quality gates; Terraform and Argo CD Gap (Terraform forbidden) |
+| P4 | Developing with AI tools and agents | Direct-ish: Kiro, Copilot, LiteLLM, custom GPTs (SW-8 limits apply) |
+| P5 | Production AI inference or agentic workloads, debugging across models, runtimes, Kubernetes, hardware | Adjacent: BS-1 |
+
+## Open Evidence Questions (asked 2026-10-22, per the recurring DevOps/SRE lesson)
+1. **Current ops at Swisscom (ART DPT):** monitoring/alerting on his Glue/Airflow pipelines (CloudWatch alarms, failure alerts, dashboards)? On-call in the current unit? Decides whether Q5/Q6 have current evidence.
+2. **Networking:** AWS VPC, security groups, private endpoints at Swisscom; firewall, DNS, proxy or load-balancer work at Bosch? Would move Q4 networking from personal-only (PP-3) to professional.
+3. **Bosch inference hardware:** did the inference containers run on GPUs? If yes, it licenses "deployed GPU-backed inference containers" only, never CUDA or GPU performance work.
+
+## Competitive Read
+- **Obvious-fit candidate:** an SRE/production engineer from a hyperscaler, GPU cloud or ML-serving company who has run vLLM/Triton-style
+  endpoints on Kubernetes at scale, written operators, and works in Terraform/Argo CD with formal SLOs and error budgets.
+- **Dennis's advantage:** production ML inference integrated into a continuously running semiconductor fab where a defect could have stopped
+  production (BS-1); Ansible depth beyond use, including extensions (BS-6); component ownership with on-call and root-cause work (SW-2);
+  current AWS IaC (CloudFormation) and Python automation; a ~10-year automation thread across five employers; current AI dev tooling.
+- **Their advantage:** inference-serving stack, GPU, Kubernetes at scale and currently, SRE formalism (SLIs/error budgets), networking.
+- **Level/scope:** NVIDIA Senior against Swisscom Staff — not a level problem; comp upside is clear.
+
+## Framing Strategy (draft — confirmed in Phase 2)
+- **Professional identity:** engineer who builds and operates the automation and delivery layer under production services, including ML
+  inference in a 24/7 environment. Never "SRE" as a title.
+- **Strongest proof points:** BS-1, BS-6, SW-2, SW-1 + CloudFormation, SW-3 (past), BS-3 SLOs, BS-4 (PoC).
+- **Explicit gaps:** inference-serving stack, GPU, current Kubernetes, SLIs/error budgets, networking, large-scale distributed systems.
+- **User directives:** none yet.
+
+### Answers received 2026-10-22 (recorded in claims.json; extraction `knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md`)
+1. Current on-duty: Opsgenie alerting, rotating engineer-on-duty role, alerts and bugs auto-create Jira tickets the team assigns and fixes → **new claim SW-11** (current).
+2. Networking: no professional configuration; automated AWS landing zones; AWS SAA cert covers AWS networking → skill `networking`, **certification context only**. Q4 stays Adjacent.
+3. Bosch inference ran on **NVIDIA GPUs** with CUDA; he did little CUDA/GPU configuration → BS-1 scope; **CUDA registered forbidden** as a skill.
+
+**Re-score after answers: Evidence Fit 65** (Recency 5→6: current on-duty; Domain 4→5: production inference on NVIDIA GPUs). Class unchanged: **Stretch** — Q5 still lacks SLIs and error budgets. Hard gate PASS.
+
+## Resume Plan (proposed 2026-10-22, awaiting approval)
+- **Profile:** International Tech, 2 pages. Contact line Bern, Switzerland, plus "EU citizen, Swiss B permit" (remote-CH role, so residency/authorization is worth one line).
+- **Summary (2-3 lines):** over ten years building and operating production software and its automation; GPU-backed ML inference integrated into a continuously running fab (BS-1); today AWS pipelines with CloudFormation and a seat in the on-duty rotation (SW-1, SW-11). No "12+ years" (template claim is open in the KB log).
+- **Skills (~5 rendered lines):** Python, SQL, Bash · AWS (Glue, Athena, S3, Airflow), CloudFormation, Ansible, Docker, Kubernetes, Linux administration, Jenkins, GitLab CI/CD · on-duty incident response (Opsgenie, Jira), SLOs, root-cause analysis, unit testing (pytest) · PySpark, Kafka, Apache Iceberg · AI dev tools: Kiro, Copilot, LiteLLM · Certs line: AWS SA Associate (2024), IBM AI Engineering (PyTorch, TensorFlow; 2020), ITIL Foundation (2016). Never: CUDA, GitOps, Terraform, SRE title, networking as a skill.
+- **Swisscom (5):** (1) SW-11 on-duty rotation, Opsgenie alerts, Jira-tracked fixes — current; (2) SW-1 + CloudFormation: his domains' ETL migrated to AWS as Python Glue jobs, IaC, Airflow — contributor to the wider programme; (3) SW-7 + SW-6 + SW-9: design and build governed data products (PySpark, Glue, Athena, Iceberg) inside the company-wide Data Mesh, unit-tested with pytest; (4) SW-3 past: built and operated Python applications on Kubernetes, deployed through automated GitLab CI/CD across multiple environments; (5) SW-2 past: owned the Fulfillment ETL components into Teradata with on-call, 2nd/3rd-level support and root-cause analysis.
+- **Bosch (4):** (1) BS-1 lead: designed and executed integration of containerized ML inference on NVIDIA GPUs into a continuously operating fab where a software defect could have halted production (Docker, Kubernetes, Ansible), reducing manual defect classification; (2) BS-6: administered and automated the Linux estate behind fab analytics and ML (Ansible from Git, run from Jenkins/GitLab), incl. the credential-caching Ansible plugin; (3) BS-3/BS-5: technical Application Owner for KLA DMS and Spotfire, defined SLOs, vendors, user training; (4) BS-4: ELK/Kafka log-management and anomaly-detection PoC with monitoring (PoC label kept).
+- **Earlier (1 each):** Fraunhofer FC-1 (Jenkins jobs for automatic SCEDAS deployment); Vizrt VZ-2 + VZ-1 (test suites as CI/CD quality gates; contributed Python to a distributed transcoding backend; no customer names); Generali GN-3 + GN-1 (web services onto a new XLDeploy deployment procedure; technical responsibility for BDD test automation in Jenkins).
+- **Total bullets:** 12.
+- **Open confirmations before generation:** (a) BS-4 — did the PoC include Grafana/Prometheus/Loki? Source is only the user's earlier CV (`thiessen_cv_master_profile.md`), not the Zeugnis or claims.json; omitted unless confirmed. (b) SW-11 — does he also write or adjust the alerts for his own pipelines? If not, the bullet says "respond to" only.
+- **Impact evidence still needed:** none quotable; every claim is metrics-unverified.
+
+## Cover Letter Decision
+- **NO.** Workday portal, no letter requested; no transition, relocation or language issue to explain; the GPU and on-duty evidence fits in the resume. The warm action (LinkedIn note) has more information value than a letter.
+
+### Confirmations received 2026-10-22 (recorded in claims.json)
+- (a) BS-4 included Grafana, Prometheus and Loki (PoC). (b) He writes alert thresholds for his own pipelines, e.g. data-quality thresholds (SW-11).
+- Generation started on the user's instruction "prepare the cv for nvidia" plus these answers; no plan objection was raised. Any plan change is an `/edit-resume` pass.
+
+## As Built (2026-10-22)
+- **Header:** headline "Staff Data & AI Engineer | Production Automation and Operations | AWS · ML Inference"; contact line Bern, Switzerland | EU citizen, Swiss B permit.
+- **Summary:** 3 rendered lines (trimmed from 4 to meet config.md's 2-3); "over ten years in production software and its automation" (measured from Generali 2015-05).
+- **Skills:** 5 categories, 8 rendered lines (trimmed from 9; cut what bullets prove: Opsgenie/Jira, ELK/Prometheus/Grafana/Loki, AWS service list, LiteLLM, ITIL). Literals for required Q3 restored after the headline cut dropped them: "infrastructure as code (CloudFormation)", "configuration management (Ansible)"; Q1 support: "PySpark (distributed processing)". PyTorch/TensorFlow only inside the IBM certificate line (validator rule). No CUDA, GitOps, Terraform, networking or SRE title.
+- **Bullets: 13** — Swisscom 5 (SW-11, SW-1, SW-7/SW-6/SW-9, SW-3 past, SW-2 past); Bosch 5 (BS-1 with NVIDIA GPUs, BS-6 estate, BS-6 plugin, BS-3/BS-5, BS-4 PoC); Fraunhofer FC-1; Vizrt VZ-2/VZ-1; Generali GN-3/GN-1. Deviation from plan: the BS-6 credential plugin became its own short bullet (Bosch 5 instead of 4) for readability.
+- **Layout:** the template's forced page break after Swisscom was removed so the Bosch NVIDIA-GPU bullet sits on page 1. Page 2 ends about half-way; no filler added.
+- **Validation:** `validate_resume_system.py --document` PASS, 0 warnings (first run failed on PyTorch without certification context; fixed). MiKTeX pdflatex: 2 pages, 0 overfull/underfull boxes. Both pages inspected visually at 80 dpi. `pdftotext -layout`: heading/employer/title/date order correct; no Capgemini.
+
+## Output Files
+- Resume: `e2e_nvidia_prod_eng_dgx_cloud_resume.tex` (source of truth) + `.pdf`
+- Cover letter: intentionally omitted (decision NO above)
+- Critique: `critique_nvidia_prod_eng_dgx_cloud.md` (2026-10-22)
+
+## Critique Summary (2026-10-22)
+- Evidence Fit: **65/100 — Stretch**
+- Hard gate: **PASS** (Q1 large-scale distributed systems is the likeliest screen rejection)
+- Document Quality: **85/100**
+- Channel Strength: **Weak** (warm LinkedIn note not yet sent)
+- Tier 1 fixes: (1) Swisscom #1 "set the alert thresholds" → "set alert thresholds". Tier 2: tense cue on the two past Swisscom bullets; one-line PP-3 personal-infrastructure entry; tighten Swisscom #1 and Bosch #1; optional "observability" on BS-4.
+
+## Status
+- Fit gate: Stretch, Evidence Fit 63 initial → 65 after the 2026-10-22 answers; hard gate PASS
+- Resume: generated and validated 2026-10-22; awaiting user review
+- Cover letter: NO
+- Critique: done 2026-10-22 — DQ 85 measured BEFORE Edit 1; **stale after Edit 1** (not re-measured; never quote a post-edit estimate as a score)
+- Application/outcome: **SUBMITTED 2026-10-22** via Workday (user-confirmed). Tracker row added (applied, stretch, 65, hard gate pass, channel weak); scout decision `applied`. Channel still Weak until the LinkedIn note is sent.
+- **Finalized 2026-10-22 on user instruction ("finalize").** Checks: session, `.tex`, `.pdf`, critique and JD present; PDF newer than `.tex`; validator PASS, 0 warnings; 2 pages; PDF metadata clean (empty Title/Author, MiKTeX pdfTeX). Submission copy `Dennis_Thiessen_Resume.pdf` created, byte-identical to `e2e_nvidia_prod_eng_dgx_cloud_resume.pdf`. No cover letter (decision NO).
+- Next: user submits via Workday → then `application_tracker.py add` (outcome applied, fit stretch, evidence-fit 65, hard-gate pass, channel weak, job-id JR2027105) and regenerate views. Not added before submission, because `add` stamps `application_date`. Warm action: LinkedIn note to an NVIDIA recruiter or DGX Cloud engineer (draft given to the user 2026-10-22). **Refresh `Dennis_Thiessen_Resume.pdf` on any recompile.**
+
+## Edit History
+
+### Edit 1 (2026-10-22): critique Tier 1 + all Tier 2, user-approved ("apply fixes and improvements as you recommend")
+- **Tier 1 #1 (truth):** Swisscom #1 "set the alert thresholds" → "set alert thresholds" (SW-11: data-quality thresholds were an example, not all thresholds).
+- **Tier 2 #1:** the two past Swisscom bullets now read "Earlier in the role, owned the Fulfillment ETL components …" and "Built and operated Python applications on Kubernetes in the same period …" (SW-2 moved before SW-3 so the two do not open identically). No dates, no unit names.
+- **Tier 2 #2:** new section **Personal Infrastructure** (PP-3): "Self-host and harden a personal Debian server, run for about ten years: nginx, a mail server, a VPN server, Docker services and Bitwarden." Personal label in the heading and the sentence; no uptime/scale figure; not presented as employer, production or datacenter work.
+- **Tier 2 #3:** Swisscom #1 tightened ("alerts" once; "resolving issues through auto-created Jira tickets"); Bosch #1 now names the GPUs once: "containerized ML inference on NVIDIA GPUs (Docker, Kubernetes, Ansible)".
+- **Tier 2 #4:** BS-4 now "Built an observability proof of concept for log management and anomaly detection on ELK and Kafka …" (PoC label kept).
+- **Verification:** validator PASS, 0 warnings; MiKTeX 2 pages, 0 overfull/underfull boxes; both pages inspected at 80 dpi (Bosch NVIDIA-GPU bullet still on page 1; Personal Infrastructure sits between Experience and Education on page 2; page 2 ends a little past half-way); `pdftotext -layout` order: Summary → Skills → Experience (Swisscom, Bosch, Fraunhofer, Vizrt, Generali) → Personal Infrastructure → Education. Bullets: 13 experience bullets unchanged in count, plus the one-line personal entry. (copy to `Dennis_Thiessen_Resume.pdf`, tracker `add`). Warm action before/after submit: LinkedIn note to an NVIDIA recruiter or DGX Cloud engineer.

resume_builder/canonical/claims.json

@@ -363,7 +363,7 @@
         "source": "Bosch Arbeitszeugnis, Dresden 31.12.2022 (primary document, read 2026-09-11); operating constraint user-clarified 2026-09-11",
         "ownership": "individual - integration and operation; model development is NOT his"
       },
-      "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. OPERATING CONSTRAINT, user-clarified 2026-09-11: the fab runs continuously, and a defect in the software could in the worst case halt production - an outcome that must never occur. That is a FAILURE-CONSEQUENCE constraint (blast radius of a bug), NOT a deployment-window constraint. Write it as 'a defect could have stopped production' or 'software whose failure would have halted a continuously operating fab'. ZEUGNIS WORDING (primary, 2026-09-11): 'Erarbeitung und Durchfuehrung von Integrationsstrategien fuer den Einsatz von Machine Learning Modellen in der 24/7 Fertigungsumgebung' - devising AND executing integration strategies for ML models in the 24/7 manufacturing environment. This corroborates the integration-and-operation framing and the 24/7 context directly. DESIGN, user-confirmed 2026-10-21 (knowledge_base/extractions/2026-10-21_software_design_architecture_clarification.md): software design and architecture of the ingestion and inference scripts were part of his Bosch work (2020-2022) and count toward a 'years of software design and architecture' requirement. Scope 'designed' to his own scripts and the inference integration - never the fab IT architecture, an architect title or design authority.",
+      "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. OPERATING CONSTRAINT, user-clarified 2026-09-11: the fab runs continuously, and a defect in the software could in the worst case halt production - an outcome that must never occur. That is a FAILURE-CONSEQUENCE constraint (blast radius of a bug), NOT a deployment-window constraint. Write it as 'a defect could have stopped production' or 'software whose failure would have halted a continuously operating fab'. ZEUGNIS WORDING (primary, 2026-09-11): 'Erarbeitung und Durchfuehrung von Integrationsstrategien fuer den Einsatz von Machine Learning Modellen in der 24/7 Fertigungsumgebung' - devising AND executing integration strategies for ML models in the 24/7 manufacturing environment. This corroborates the integration-and-operation framing and the 24/7 context directly. DESIGN, user-confirmed 2026-10-21 (knowledge_base/extractions/2026-10-21_software_design_architecture_clarification.md): software design and architecture of the ingestion and inference scripts were part of his Bosch work (2020-2022) and count toward a 'years of software design and architecture' requirement. Scope 'designed' to his own scripts and the inference integration - never the fab IT architecture, an architect title or design authority. HARDWARE, user-confirmed 2026-10-22 (knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md): the inference ran on NVIDIA GPUs, with CUDA used for the performance boost - his words: 'The ML inference did ran on NVIDIA GPUs (using CUDA for performance boost - but i never did much configuration there)'. Writable: 'GPU-backed inference containers on NVIDIA GPUs' as a description of what he integrated and deployed. CUDA is a property of the system, not his skill: he did little CUDA or GPU configuration, so never claim CUDA work, GPU configuration, GPU tuning or GPU performance analysis.",
       "allowed_verbs": [
         "integrated",
         "containerized",
@@ -375,7 +375,12 @@
         "trained the image classification model",
         "owned the full ML lifecycle",
         "a production line that could not be stopped for a deployment",
-        "could not be stopped for a deployment"
+        "could not be stopped for a deployment",
+        "CUDA programming",
+        "configured CUDA",
+        "GPU performance tuning",
+        "optimized GPU performance",
+        "GPU performance analysis"
       ],
       "metrics": "qualitative reduction in manual classification is source-backed; exact amount unverified"
     },
@@ -414,7 +419,7 @@
         "source": "Bosch Arbeitszeugnis, Dresden 31.12.2022 (primary document, read 2026-09-11)",
         "ownership": "individual - proof of concept"
       },
-      "scope": "Built an ELK/Kafka anomaly-detection proof of concept and monitoring; not a company-wide observability platform. ZEUGNIS WORDING (primary): 'Durchfuehrung eines POCs fuer das Logfile-Management mit ELK-Stack' - the PoC framing is confirmed by the document itself, not just by policy.",
+      "scope": "Built an ELK/Kafka anomaly-detection proof of concept and monitoring; not a company-wide observability platform. ZEUGNIS WORDING (primary): 'Durchfuehrung eines POCs fuer das Logfile-Management mit ELK-Stack' - the PoC framing is confirmed by the document itself, not just by policy. STACK, user-confirmed 2026-10-22 (knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md): besides ELK and Kafka, the PoC included monitoring and alerting with GRAFANA, PROMETHEUS and LOKI. That was previously recorded only in the user's earlier CV (thiessen_cv_master_profile.md), not in the Zeugnis, which names the ELK log-management PoC. All of it stays PoC-scoped: never a production monitoring platform.",
       "allowed_verbs": [
         "built",
         "implemented",
@@ -819,6 +824,36 @@
         "formal mentoring programme - not established"
       ],
       "metrics": "none; number of junior developers and outcomes are not recorded"
+    },
+    {
+      "id": "SW-11",
+      "evidence": {
+        "source": "user confirmation 2026-10-22; knowledge_base/extractions/2026-10-22_oncall_networking_gpu_clarification.md",
+        "ownership": "team-role - member of a rotating engineer-on-duty role; individual - alert thresholds for his own pipelines"
+      },
+      "scope": "CURRENT on-duty and alert handling at Swisscom (user answer 2026-10-22 to 'do you monitor or get alerted on your Glue/Airflow pipelines; is there on-call in the current unit?'). Verbatim: 'Yes, we get alerted ... we are using opsgenie and we have a engineer on duty role which rotates. Bugs and Alerts also automatically create Jira issue tickets which we dann asign an fix.' Established: alerting through OPSGENIE; a ROTATING ENGINEER-ON-DUTY role in his team, which he takes part in; bugs and alerts automatically create Jira tickets that the team assigns and fixes. PHASE: current (asked about the current unit, ART DPT); never date it. This is the current counterpart of the past SW-2 on-call (Rufbereitschaft) at ART B2B. NOT established: who built or configured the alerting, the Opsgenie routing or the Jira automation (do not claim he did); which metrics or logs feed the alerts; SLIs, SLOs or error budgets; any incident count, response time, MTTR or availability figure; whether the duty covers nights and weekends (do not write 24/7). ALERT DEFINITIONS, user-confirmed 2026-10-22 (follow-up 'do you also write or adjust the alerts for your own pipelines?'): 'yes, i wrote e.g. thresholds for data quality etc....' So he defines alert thresholds for his own pipelines, data-quality thresholds being the named example. Writable: 'defined alert thresholds, including data-quality checks, for my pipelines'. Still NOT established: other alert types (latency, freshness, cost) - do not list them; the Opsgenie routing and the Jira automation, which he did not set up.",
+      "allowed_verbs": [
+        "serve",
+        "take part in",
+        "respond to",
+        "triage",
+        "fix",
+        "resolve",
+        "define",
+        "set"
+      ],
+      "forbidden": [
+        "configured Opsgenie",
+        "set up Opsgenie",
+        "designed the on-call process",
+        "owned incident management",
+        "SRE role or title",
+        "24/7 on-call",
+        "any incident count, MTTR, response-time or availability figure",
+        "built the alerting platform",
+        "set up the Opsgenie or Jira integration"
+      ],
+      "metrics": "none"
     }
   ],
   "skills": [
@@ -1273,6 +1308,53 @@
       "evidence": "production-historical",
       "output": "allowed",
       "note": "Registered 2026-10-14 from existing canonical claim scopes - no new user statement, same procedure as Jenkins and GitLab CI/CD on 2026-10-08. Was evidenced in scopes but absent from this skills list, so a skills line could not cite it (surfaced by the BIS jr100498 package). BS-2 (Bosch 2020-2022): data services over Hadoop/Impala for fab analysis teams. Query and data-service use; does not establish Hadoop cluster administration or Cloudera administration."
+    },
+    {
+      "name": "Opsgenie",
+      "evidence": "hands-on-current",
+      "output": "allowed-with-context",
+      "context": "SW-11: alerting for his team's rotating engineer-on-duty role at Swisscom.",
+      "note": "User-confirmed 2026-10-22. Use as a responder, never as its administrator or the author of the alert routing."
+    },
+    {
+      "name": "CUDA",
+      "evidence": "system-context-only",
+      "output": "forbidden",
+      "note": "User-confirmed 2026-10-22: the Bosch inference (BS-1) ran on NVIDIA GPUs with CUDA for the performance boost, but he 'never did much configuration there'. CUDA describes that system, not a skill he has. Registered forbidden so the scanner catches it on documents; the honest wording is 'GPU-backed inference on NVIDIA GPUs' (BS-1). In an interview he can say the models used CUDA and that his part was integration and deployment."
+    },
+    {
+      "name": "networking",
+      "evidence": "certification-and-personal",
+      "output": "certification-context-only",
+      "note": "User answer 2026-10-22 to 'any AWS VPC, security-group or private-endpoint work at Swisscom, or firewall, DNS, proxy or load-balancer work at Bosch?': 'No, not really - probably only my AWS solution architect cert covers AWS network topics... but day to day work is using automated landing zones.' So: no professional network configuration. AWS networking is covered by the AWS Certified Solutions Architect - Associate (26 Sep 2024, active; thiessen_certifications.md #7), certification context only. At work he USES automated AWS landing zones provided by the platform - never claim he built or designed landing zones, VPCs or network topology. Personal: PP-3 runs nginx, a VPN server and a mail server (personal label required; PP-3 forbids datacenter or physical-infrastructure claims). JD lines on networking fundamentals stay Adjacent at best."
+    },
+    {
+      "name": "ELK Stack",
+      "evidence": "proof-of-concept",
+      "output": "allowed-with-context",
+      "context": "BS-4: Bosch log-management and anomaly-detection proof of concept (2020-2022).",
+      "note": "Registered 2026-10-22. ELK named in the Bosch Zeugnis. Always keep the PoC label; never imply production or platform ownership."
+    },
+    {
+      "name": "Grafana",
+      "evidence": "proof-of-concept",
+      "output": "allowed-with-context",
+      "context": "BS-4: Bosch log-management and anomaly-detection proof of concept (2020-2022).",
+      "note": "Registered 2026-10-22. User-confirmed 2026-10-22; previously only in the user's earlier CV. Always keep the PoC label; never imply production or platform ownership."
+    },
+    {
+      "name": "Prometheus",
+      "evidence": "proof-of-concept",
+      "output": "allowed-with-context",
+      "context": "BS-4: Bosch log-management and anomaly-detection proof of concept (2020-2022).",
+      "note": "Registered 2026-10-22. User-confirmed 2026-10-22; previously only in the user's earlier CV. Always keep the PoC label; never imply production or platform ownership."
+    },
+    {
+      "name": "Loki",
+      "evidence": "proof-of-concept",
+      "output": "allowed-with-context",
+      "context": "BS-4: Bosch log-management and anomaly-detection proof of concept (2020-2022).",
+      "note": "Registered 2026-10-22. User-confirmed 2026-10-22; previously only in the user's earlier CV. Always keep the PoC label; never imply production or platform ownership."
     }
   ],
   "global_forbidden_output_patterns": [

resume_builder/reference/current_applications.md

@@ -5,7 +5,7 @@ Use current outcome and next action below. Imported legacy notes are historical,
 
 | Application | Current outcome | Interview reached | Next action |
 |---|---|---|---|
-| NVIDIA - Senior DevOps Engineer, Zürich (req JR2024256) | applied | not recorded | Await response. At first contact, clarify homeoffice/onsite requirements and total package. Baseline is 3 homeoffice days/week; fewer days require a clear overall improvement. Note: `claims.json` has no shell-scripting entry, so **Bash is omitted** from Skills although the JD names it and ten years of Debian admin plus Ansible work plainly involve it - same shape as the Ansible/Linux gaps found 2026-08-27. Confirm with the user and record it. Groovy stays out regardless. |
+| NVIDIA - Senior DevOps Engineer, Zürich (req JR2024256) | applied | not recorded | No action expected; req pulled ~2026-09-20. Treat as dormant. Does not block NVIDIA JR2027105 (Senior Production Engineer, DGX Cloud). |
 | RUAG C5I - AI Engineer C5I, Thun (application ID 18027) | rejected_after_interview | yes | Abgeschlossen: Absage nach Interview; Anforderungen vor allem beim Homeoffice nicht vereinbar. Keine weitere Aktion fuer diese Bewerbung offen. |
 | Schweizer Armee - Kommando Cyber (Kdo Cy), DevOps Engineer III (Data Platform), Zimmerwald (Ref JRQ$540-19848) | rejected_no_interview | not recorded | Closed; no pending application action. Read recorded history before reopening. |
 | Citadel Securities — Platform Engineer, Research Platform Engineering, Zürich (also NY/Miami) | applied | not recorded | Await response. At first contact, confirm whether the application is for the Zurich seat; relocation to the US remains ruled out. Clarify homeoffice/onsite requirements and the Zurich compensation package. Baseline is 3 homeoffice days/week; fewer days require a clear overall improvement. The posted US salary range is not a Zurich figure. |
@@ -36,6 +36,7 @@ Use current outcome and next action below. Imported legacy notes are historical,
 | Kongsberg Discovery (Seatex) — Lead Engineer, DevOps and Test Infrastructure, Trondheim | applied | not recorded | Wait for Kongsberg response. Screen prep: embedded targets/HIL question, recency of DevOps/test work (2015-22), relocation to Trondheim, clearance eligibility as German citizen resident in Switzerland, on-site expectations. Package: output/Kongsberg_Lead_Engineer_DevOps_Test_Infra/. |
 | BIS (Bank for International Settlements) — Senior/Lead Data & Analytics Engineer, MED IT, Basel (jr100498, 3-yr term) | applied | not recorded | Await BIS response. If contacted: ask office days and total package; weigh the 3-year fixed term against the permanent Swisscom post. Interview prep is in the critique: SDMX answer, Swisscom dashboard tool (not in KB), attributed schedule example for project planning, Lead story. |
 | Google — Staff Software Engineer, Shopping Graph (Feeds Infra), Zürich | applied | not recorded | Await Google response. Google limit: next Google application window counts this submit (3 per rolling 30 days). At first contact clarify level (L6 vs L5), Zurich hybrid days vs current 3 homeoffice days, and total package. Interview prep: SW-4 prototype story as the technical-direction example, plus one concrete Staff-level example the user names; Java currency; end-to-end system design with scale and trade-offs. Package: output/Google_Staff_SWE_Shopping_Graph/. |
+| NVIDIA — Senior Production Engineer, DGX Cloud, Switzerland Remote (JR2027105) | applied | not recorded | Await response. Send the LinkedIn note to an NVIDIA recruiter or DGX Cloud engineer (draft in session). Interview prep: incident debugged end to end (SW-2/SW-11); safe model rollout in the fab (BS-1, rollout mechanics not in KB); Kubernetes recency; distributed systems at scale; vLLM/SGLang/Dynamo not used. Package: output/NVIDIA_ProdEng_DGX_Cloud/. |
 
 ## Current targeting notes