claude-resume-kit/output/Microsoft_ISE_Senior_SWE/session_microsoft_ise.md

Session: Microsoft Senior Software Engineer — Industry Solutions Engineering (ISE), Zürich

JD Info

• File: job_scout/jd_microsoft_ise.txt (copy: output/Microsoft_ISE_Senior_SWE/JD_microsoft_ise.txt)
• JD source: live fetch 2026-07-03 via Eightfold PCSX API (apply.careers.microsoft.com, req 200040836), verbatim
• Role: Senior Software Engineer, Industry Solutions Engineering (ISE) — IC4, Software Engineering profession
• Company: Microsoft (ISE = global non-billing co-engineering org inside MCAPS; engineers code side-by-side with strategic customers)
• Bundle: HYBRID — primary bundle_ml_ai_engineer.md (JD's decisive axis: production AI systems, LLM/RAG, model eval/monitoring), secondary bundle_data_engineer.md (strongest evidence; reframing map for bridging bullets)
• Format: Resume (2-page, resume.cls) + 1-page cover letter
• Salary/Details: Published base band CHF 146,200–245,900 (IC4, Switzerland). Zürich. Travel up to 25%. German "beneficial". Posted ~2026-06-26, "open minimum 5 days, ongoing until filled" — apply promptly.

JD Analysis

Requirements

#	Requirement	Match	Evidence
1	BS in CS/related + 4+ yrs engineering w/ coding (C, C++, C#, Java, JavaScript, Python)	Direct	B.Eng. + M.Eng. (Software Design & Engineering); 12+ yrs; Python + Java strong (per feedback_cpp_emphasis: lead Python/Java, do NOT oversell C++; C#/.NET minor via Fraunhofer — secondary mention only)
2	Familiarity with deploying and operating AI systems in production environments	Bridge (high)	Swisscom: governed data products + metadata mgmt on AWS as the data foundation for company agentic-AI programme (SW-7); operates production pipelines feeding AI workloads; LiteLLM gateway API usage. NOT model-serving ownership — hedge verbs
3	Experience building or integrating AI/ML or LLM-based solutions, prompt engineering, RAG	Bridge (medium-high)	Verified toolchain ONLY: Kiro (spec-driven AI dev), VS Code + Copilot, LiteLLM (created/used LLM API integrations), custom GPTs fed with domain knowledge (prompt engineering + grounding = RAG-adjacent). NEVER LangChain/LangGraph/LlamaIndex (config ban)
4	Understanding of model evaluation, data quality, and performance monitoring	Bridge (high)	Data quality is core Swisscom work (data products with contracts/metadata); monitoring/observability from DevOps history (Bosch/Swisscom pipelines). Model evaluation specifically = thin — claim "data quality & performance monitoring" side confidently, model-eval only as understanding
5	Design/deliver solutions with modern software engineering practices + cloud technologies	Direct	AWS-heavy platform work (Swisscom), CI/CD, IaC-adjacent DevOps at Bosch/Swisscom; Azure = gap (AWS-primary — frame cloud skills as transferable, name AWS honestly)
6	Customer-facing co-engineering, cross-functional teams (SWE + data scientists + TPMs + designers)	Direct	Bosch: co-owned Spotfire analytics platform for internal customers, co-presented TIBCO Analytics Forum 2022; Swisscom: onboarding domain teams onto data products, guiding producers; Fraunhofer: applied research with industry partners
7	DRI/on-call: monitor systems, restore service, playbooks	Bridge (high)	Application Owner / Component Owner roles = production ownership incl. incident response for his components (scope-disciplined: HIS components, not org-wide)
8	Open source contribution / variety of technologies "not just Microsoft"	Bridge (medium)	Polyglot record (Python/Java/C#, AWS/on-prem, Spotfire/TIBCO); personal Solidity work. No notable OSS contributions — do not claim
9	Travel up to 25%	Direct	Explicitly matches mobility appetite (user_international_mobility: travel-OK from Bern, cross-cultural work history NO/DE/CH/Shanghai)
10	German language beneficial (preferred qual)	Direct	Native German + fluent English — differentiator for Swiss/DACH customer engagements

ATS Keywords

• ML/AI: AI systems in production, LLM, RAG, prompt engineering, model evaluation, AI/ML solutions, Copilot, agentic
• Domain: co-engineering, customer engagements, industry solutions, cloud solutions, open source
• Methods: modern software engineering practices, CI/CD, code review, design documents, observability, monitoring, incident response (DRI), playbooks, reliability, performance, maintainability
• Tools: Python, Java, C#, cloud (AWS→Azure bridge), Kubernetes/containers, Git/GitHub
• Soft skills: cross-functional collaboration, stakeholder requirements, growth mindset, mentoring/guiding engineers

Gap Assessment

• Direct: coding depth (Python/Java, 12+ yrs), cloud data/platform engineering, customer-facing co-engineering, cross-functional work, German, travel appetite
• Bridge (confidence): production AI operations via data-foundation-for-AI (high); LLM integration via LiteLLM/custom GPTs/Kiro (medium-high — verified tools only); DRI/on-call via Application/Component Owner (high); model-eval understanding (medium — hedge as "understanding")
• Gap (cannot claim): Azure-specific services (AWS-primary — say so honestly), model training/fine-tuning ownership, notable OSS contributions, C++ depth, JavaScript/front-end depth
• Fabrication tripwires: NO LangChain/LangGraph (config ban); no "built the Data Mesh" (scope discipline); no claiming model-serving/MLOps platform ownership

Company Context

• Mission: Empower every person and organization to achieve more. ISE specifically: non-billing global engineering org (inside MCAPS) that co-develops production code side-by-side with strategic customers' engineers, then feeds learnings back into Microsoft products. Public Engineering Fundamentals Playbook (microsoft.github.io/code-with-engineering-playbook) codifies how they work: agile ceremonies, code-with, testing, observability fundamentals.
• This role: Zürich-based Senior SWE (IC4) embedded in customer engagements — likely Swiss/DACH strategic accounts (banks, pharma, industrials) given the German-beneficial flag. Recent ISE publication themes: enterprise RAG with document-permission propagation (Entra ID → AI Search), coordinator-based multi-agent architectures in production, vector/hybrid search evaluation frameworks, multimodal RAG fine-tuning experiments.
• Culture: "Meet customers where they are — their languages, their frameworks, their OS." Explicitly not-just-Microsoft tech. Growth mindset, informal/flexible, travel ~25%. Cross-functional pods (SWE + DS + TPM + design).
• Swiss context (CL-relevant): Microsoft committed USD 400M (June 2025) to expand Swiss datacenter capacity near Zürich and Geneva with in-country data residency for 50k+ customers; AI Tour Zürich April 2026 (3,000+ attendees); Swiss AI Tech Accelerator cohort 3 (Jan 2026).
• "Why them" angle: ISE is the rare role where a platform/data engineer with enterprise-scale AI-foundation experience gets to do hands-on engineering ACROSS companies instead of inside one — Dennis's multi-country, multi-industry ramp (Fraunhofer research → Bosch automotive → Generali insurance → Swisscom telecom) is exactly the "walk into a new enterprise and be productive fast" profile ISE hires for.

Framing Strategy

• Lead narrative: Staff-level data & AI platform engineer (12+ yrs, telecom/automotive/research) who builds the governed data and platform foundations that make production AI work inside large enterprises — now bringing that enterprise-hardened, customer-embedded engineering to ISE's co-engineering model.
• Reframing map:
  • "data products / Data Mesh contribution (SW-7 agentic-AI foundation)" → "data foundations for production AI / agentic systems" (scope-disciplined)
  • "LiteLLM gateway APIs + custom GPTs with domain grounding" → "integrating LLM-based solutions; prompt engineering and knowledge grounding" (verbatim-verified tools only)
  • "Application Owner / Component Owner incident duty" → "DRI-style production ownership: monitoring, incident response, restoration playbooks" (his components only)
  • "Bosch Spotfire platform co-ownership + TAF 2022 co-presentation" → "customer-facing platform engineering and technical evangelism with internal customers"
  • "AWS Glue/Athena/Redshift stack" → "cloud-native engineering (AWS; concepts transferable to Azure)" — name AWS, never claim Azure depth
• Emphasize: Python/Java polyglot depth; enterprise data/AI platform work; customer-embedded collaboration (Bosch analytics platform, Swisscom domain onboarding); production ownership/reliability; German+English; cross-industry adaptability (4 industries, 3 countries + Shanghai)
• Downplay: C++ (per feedback_cpp_emphasis), C#/.NET (one mention max), Spotfire/BI tooling specifics, security-champion badge (JD doesn't gate on security)
• CL hooks: (1) ISE Engineering Fundamentals Playbook — "code-with" model matches how he onboards domain teams onto data products; (2) ISE blog's enterprise-RAG/permission-propagation work — mirrors his governed-data-products + metadata/access-control reality at Swisscom; (3) Microsoft's CHF 400M Swiss datacenter/data-residency investment — he builds exactly the kind of regulated, in-country data platforms Swiss customers will bring to ISE; (4) German-native for DACH accounts + 25% travel appetite.
• User directives: None given (autonomous Friday-run build). Defaults from hybrid priority matrix.

Critique Context

• Reviewer persona: ISE engineering manager or senior/principal SWE in EMEA doing hiring-manager screen. Reads 50+ CVs per req. Daily work: scoping customer engagements, unblocking pods, code reviews. Impressed by: production evidence, breadth across stacks, customer-facing engineering signals, crisp scope-honest claims. Bored/annoyed by: buzzword AI claims without operational substance, tool soup, solo-hero claims over org-scale objects.
• Competitive landscape: Other applicants = senior SWEs from consultancies (Accenture/Avanade), Azure-native engineers, ex-FAANG generalists. The "obvious fit" has Azure depth + OSS visibility. Dennis's edge: genuine enterprise data-platform depth + regulated-industry scars + German + already-Swiss (no visa/relocation friction — EU/EFTA citizen in Bern).
• Domain vocabulary: "code-with" / co-engineering, engagement, crew/pod, engineering fundamentals, DRI, game days, grounding, RAG, agentic workflows, Foundry/Copilot ecosystem (use sparingly — only where honest).

Cover Letter Plan

• Institution type: Industry (big tech, customer-facing engineering org)
• Paragraph count: 4 paragraphs, 250–300 words, 1 page
• P1 hook: ISE's co-engineering model ("code-with", Engineering Fundamentals Playbook) + why an enterprise data/AI platform engineer wants to do it across customers; Zürich req + German-beneficial fits him natively
• P2 evidence: Swisscom — governed data products on AWS as the data foundation for the company-wide agentic-AI programme (scope-disciplined); LLM integration via LiteLLM APIs + domain-grounded custom GPTs; production ownership (Application Owner, incident response)
• P3 evidence: cross-industry, cross-country ramp record (Fraunhofer → Bosch → Generali → Swisscom; NO/DE/CH + Shanghai) = productive-fast-in-new-enterprise; Bosch: co-owned analytics platform for internal customers, co-presented at TIBCO Analytics Forum 2022
• Domain pivot: "The foundations that make enterprise AI actually work — governed data, metadata, reliable pipelines — are what I build; ISE is where that work meets customers directly."
• Jargon level: Technical (hiring-manager-readable, not HR-safe fluff)
• "Why them" hook: Microsoft's USD 400M Swiss datacenter/data-residency expansion → the regulated Swiss enterprises he knows from the inside are exactly ISE Zürich's customer base
• Hook verification (2026-07-03, all VERIFIED):
  • Engineering Fundamentals Playbook / "code-with" → https://microsoft.github.io/code-with-engineering-playbook/ (GitHub repo tagline: "the playbook for 'code-with' customer or partner engagements")
  • ISE blog enterprise RAG / permission propagation → https://devblogs.microsoft.com/ise/sharepoint-doc-level-access/ ("Propagating SharePoint Document Permissions to AI Search and RAG Pipelines")
  • USD 400M Swiss datacenter expansion, in-country data residency → announced 2025-06-02; https://www.itpro.com/infrastructure/data-centres/microsoft-invests-usd400-million-to-expand-swiss-data-centers

Bullet Plan

Calibration: sent Google DE resume = 17 bullets (SW 6, BS 4, FC 2, VZ 2, GN 3), 2 pages clean. Recommended here: 16 confirmed + 2 fillers pending Page Fill Gate.

Position 1 — Swisscom · Staff Data, Analytics & AI Engineer (6 bullets, 12 lines)

#	ID	Achievement (ISE framing)	Variant	JD Match
1	SW-7	LEAD — governed data products + active metadata on AWS within Swisscom's company-wide Data Mesh (scoped verb per feedback_swisscom_datamesh_ownership) — the grounded-retrieval data foundation downstream AI/agentic workflows query	2L	Req 2+3 (AI systems in production, RAG/grounding) — Bridge high
2	SW-3	Python apps on Kubernetes + GitLab CI/CD — containerized, ML-ready delivery, agile DevOps	2L	Req 1+5 Direct
3	SW-1	AWS migration of legacy Teradata/Oracle ETL → S3/Glue/Athena+Iceberg/Redshift/Airflow/CloudFormation	2L	Req 5 Direct (cloud)
4	SW-2	Component Owner, business-critical Fulfillment ETL — on-call SLA, governance, data quality	2L	Req 7 (DRI) Bridge high + Req 4 (data quality)
5	SW-4	B2B data products/dashboards, stakeholder partnership, root-cause analysis under 2nd/3rd-level support	2L	Req 6 Direct (cross-functional, customer-facing)
6	SW-6	PySpark distributed processing (scale signal)	2L	Req 5 supporting

x SW-5 Security Champion — FORCED EXCLUSION (memory: team role not award, only when JD gates on security; JD doesn't)

Position 2 — Bosch · title "(Senior) Data & ML Engineer" per title-flexibility rule (4 bullets, 8 lines)

#	ID	Achievement (ISE framing)	Variant	JD Match
1	BS-1	LEAD — containerized ML inference (Docker/K8s/Ansible) into 24/7 semiconductor production; automated image-based defect classification	2L	Req 2 DIRECT ("deploying and operating AI systems in production") — flagship
2	BS-3+BS-5	Application Owner + co-owned Spotfire analytics platform for internal customers — SLOs, C# extensions, training, vendor mgmt (TAF 2022 talk reserved for CL)	2L	Req 6 Direct (customer-facing platform eng) + Req 7
3	BS-4	ELK+Kafka anomaly-detection PoC with Grafana/Prometheus/Loki	2L	Req 4 (performance monitoring/observability)
4	BS-2	Multi-language data services (Python/Java/C#) over OracleDB + Hadoop/ImpalaSQL	2L	Req 1 Direct (polyglot: 3 of the JD's 6 languages)

Position 3 — Fraunhofer · Research Software Engineer (2 bullets, 4 lines)

#	ID	Achievement	Variant	JD Match
1	FC-2	"Contributed" ML/NLP components to ARTUS sea-rescue transcription research (hedged verb MANDATORY)	2L	Req 3 supporting (applied ML)
2	FC-1	SCEDAS C#/.NET development + independently established Jenkins CI/CD with quality gates	2L	Req 1 (C#) + Req 5 (modern practices, initiative)

o FC-3 MISSION microservices (Express.js/JavaScript/Docker) — FILLER #1 if Page Fill Gate needs it (JavaScript checkbox)

Position 4 — Vizrt · DevOps Engineer, Bergen NO (2 bullets, 4 lines)

#	ID	Achievement	Variant	JD Match
1	VZ-1	Python/C++ distributed video transcoding backend (CNN/BBC/Al Jazeera scale)	2L	Req 1 (languages, distributed systems) + intl breadth
2	VZ-2	A/V test automation + CI/CD quality gates integration	2L	Req 5 (engineering fundamentals — ISE playbook resonance)

Position 5 — Generali GDIS · Software Engineer → IT Consultant (2 bullets, 4 lines)

#	ID	Achievement	Variant	JD Match
1	GN-1	Introduced BDD + technical ownership + Java Community evangelism/training	2L	Req 6 (knowledge sharing/mentoring — strong ISE culture fit)
2	GN-3	Java/J2EE workflow-portal features, XLDeploy migration, Camel/Spring Boot PoC	2L	Req 1 (Java checkbox early career)

o GN-2 UIPath RPA — FILLER #2 (automation breadth; weakest) x CA-1 Capgemini — FORCED EXCLUSION (user preference: never list, 6-month stay)

Budget: 16 recommended 2L bullets (32 rendered lines) + up to 2 fillers → 16–18 vs proven 17-bullet 2-page layout. PASS range. Skills plan (4-3-2-2-2): (1) Programming & Data Engineering, (2) Cloud & Infrastructure (AWS + K8s/Docker), (3) AI/ML & GenAI tooling — incl. memory-verified LiteLLM, GitHub Copilot, custom GPTs, Kiro (NOT in taxonomy — flagged; NEVER LangChain), (4) DevOps & Observability, (5) Certifications (AWS SAA, IBM AI Engineering, Udacity). Summary headline (draft): "Staff Data & AI Engineer | Python · Java · AWS · Kubernetes | Production AI Foundations & Customer-Facing Platform Engineering" Focus directive impact: none given — hybrid priority-matrix defaults (ML/AI primary, DE secondary).

Output Files

• Resume: output/Microsoft_ISE_Senior_SWE/e2e_microsoft_ise_resume.tex
• Cover Letter: output/Microsoft_ISE_Senior_SWE/e2e_microsoft_ise_cover_letter.tex
• Critique: output/Microsoft_ISE_Senior_SWE/critique_microsoft_ise.md

Status

• Phase 0: DONE (confirmed by user 2026-07-03)
• Phase 1: DONE (16 bullets confirmed 2026-07-03 + fillers FC-3/GN-2 authorized for Page Fill Gate; headline + skills plan approved as drafted)
• Phase 2 Resume: DONE (2026-07-03)
  • Summary: DONE (543 rendered chars, 5 lines, orphan OK)
  • Skills: DONE (4-3-2-2-2, 13 dashes; AI/ML group carries memory-verified GenAI toolchain: LiteLLM, custom GPTs w/ grounding (RAG), prompt engineering, Copilot, Kiro — NO LangChain)
  • Swisscom (6 bullets): DONE — SW-7 lead scoped ("within Swisscom's company-wide Data Mesh")
  • Bosch (4 bullets): DONE — title "(Senior) Data & ML Engineer"; BS-4 honest "proof of concept" restored (Google version had dropped it)
  • Fraunhofer (3 bullets): DONE — FC-2 "Contributed" hedge; filler FC-3 added at Page Fill Gate
  • Vizrt (2 bullets): DONE
  • Generali (3 bullets): DONE — filler GN-2 added at Page Fill Gate
  • Compile: DONE — 2 pages, MiKTeX clean, 18 variable bullets all 189–210 rendered chars (max 218, zero OVER)
  • Page fill: page 2 fuller than sent Google baseline by one 2L bullet (baseline scored 85.5, cleared recruiter screen); strict ≤3-line rule not achievable without KB-unsupported padding
  • AI fingerprint scan: banned words CLEAN, 0 rendered em-dashes, no vague -ing endings, dates consistent
• Cover Letter: DONE (2026-07-03)
  • 4 paragraphs, 299 words, 1 page, moderncv, MiKTeX clean compile (T1 fontenc + microtype expansion=false fix, same as Google CL)
  • Hooks: Playbook "code-with" (P1), ISE SharePoint-permissions-to-RAG blog (P2), USD 400M Swiss datacenter/data-residency (P4) — all web-verified, sources in Cover Letter Plan
  • Em-dashes: 0; banned-word scan clean; no generic opener; TAF 2022 talk used in P3 as planned (reserved for CL); AWS named honestly (SAA), no Azure claim
  • All claims traceable to resume bullets or verified memory; awaiting user approval at /make-cl STOP
• Critique: CURRENT (Pass 2, 2026-07-03) — 85.8/100 → critique_microsoft_ise.md
• APPROVED + FINALIZED 2026-07-03. Submission PDFs: Dennis_Thiessen_Resume.pdf (2pp) + Dennis_Thiessen_Cover_Letter.pdf (1pp), verified identical to latest compiles. Package complete (12 files + 2 submission copies).
• SENT 2026-07-03 — application submitted at apply.careers.microsoft.com (req 200040836). Logged in CLAUDE.md Active Sessions + job_scout decisions.json (applied, 85.8/100).
• Next: await Microsoft recruiter response

Critique Summary (Pass 2, 2026-07-03)

• Score: 85.8/100 (Pass 1: 83.3 → Tier 1+2 applied same day, user-directed). ATS 20/20, truthfulness PASS, AI-fingerprint CLEAN, JD integrity PASS. In the 85+ submit band.
• Applied fixes (resume only, CL untouched): (1) "model evaluation" added to AI & ML skills line — JD RQ triple now verbatim; (2) "Open-source stack:" label on Kafka/Airflow/Spark line; (3) summary rewrite: +LLM, +cross-functional, on-call bound to Component Owner (precision), tail "platform co-ownership, workshops, training" dropped for 5-line budget (signal retained in BS-3/GN-1)
• NOT applied: conditional metric for BS-1/SW-1 — KB has no verified numbers beyond 24/7 and 300mm; adding one would fabricate
• Compile verified: 2pp, summary 5 lines no orphan, skills lines single-line, page break matches baseline
• Interview likelihood: ATS 95% / Recruiter 80% / HR 80% / HM 65% / Panel 70%. Max ≈ 86.5; hard ceiling ≈ 88 (Azure, OSS contributions — not resume-editable)
• CL: PASS all 6 sub-checks — hooks verified same-day with URLs
• Known exception: page-2 bottom whitespace ~1/3 page (documented, matches sent Google baseline; padding would violate anti-fabrication)