~/programmiere

Matthias
Nordwig

Fifteen years of turning person-year problems in banking into shipped systems. Lately at speedrun pace.

Technical lead · AI systems in regulated finance

Email meLinkedIn
Career%attempt 1 · no resets
  1. entory AGConsultant1y
  2. BearingPointConsultant → Manager5y
  3. KfWProject Manager, Inhouse Consulting2y
  4. KfWSenior Manager, Compliance4y
  5. KfWSenior Manager (Technical Lead), Data Analytics & AIrunning
  6. Matrixside run · offline RAG app14d
Personal best14d 00:00:00from 48 h proof of concept to production app

Boss fights

Every station, the biggest thing it threw at me. Health bars are scope in person-days — the largest was 6,000.

2006–2007

entory AG Consultant

  • Critically escalated data-warehouse project

    Stepped up as technical lead, restored client trust, delivered.

  • IFRS KPIs that kept changing

    Self-service ETL so business units model KPIs without blocking engineering.

2007–2012

BearingPoint Consultant → Manager

  • Credit early-warning system

    Designed and delivered; scaled from HVB to the wider UniCredit group.

  • SEPA Direct Debit rollout at a Landesbank

    Led data architecture and integration.

  • FSA liquidity architecture after 6 months of gridlockgold split

    Shipped in 3 months by reverse-engineering a proprietary data mart.

  • An empty practice

    Built the Risk Management & Early Warning practice from scratch.

2012–2014

KfW Project Manager, Inhouse Consulting

  • Getting a compliance transformation approved

    Project and financial steering framework that won executive approval.

  • Reputational risk & KYC

    Target operating models for the C-level.

2014–2018

KfW Senior Manager, Compliance

  • Horizon scanning meets impact assessment

    Automated workflow system connecting both.

  • One compliance process, 100+ international offices

    Designed the global process.

  • EBA, Basel, FSB

    Analytical frameworks for regulatory impact assessments.

2018–now

KfW Senior Manager (Technical Lead), Data Analytics & AI

  • No GenAI capability yet

    Bootstrapped an internal GenAI R&D function and took prototypes to production.

  • Regulatory content nobody can search

    Architected and built an audit-ready RAG platform — design, backend, compliance documentation.

  • Complex data integrationsgold split

    Prototyped an autonomous SQL IDE; a typical integration down to 2 days.

  • Enterprise delivery

    Led end-to-end delivery at up to 6,000 person-days of scope.

Side quests

Built after hours, mostly with AI agents I direct and verify.

2 weeks

Matrix

Production-grade offline RAG app — pure-Rust core, Tauri, runs on macOS and iPhone. On-device ONNX and GGUF inference, layout-aware PDF chunking, GraphRAG ontology extraction.

Scaled from a 48-hour proof of concept.

~4,000 installs

GDPR Compliant ReCaptcha

WordPress spam protection without Google: proof of work in the browser instead of tracking.

Maintained by an autonomous loop with a deterministic approval gate.

Founder

Whistle-Blower.net

Privacy-preserving B2B compliance SaaS with a zero-dependency architecture and password-less flows.

No third-party services — full data control by design.

Loadout

AI & GenAI
RAG pipelines, embeddings & semantic search, agentic workflows, on-device AI (ONNX, GGUF), human-in-the-loop systems
Engineering
Rust, Python, TypeScript, LangChain / LangGraph, MLflow, PGVector, FAISS, DuckDB, SQL & data modeling
Architecture
hybrid & offline-first systems, cross-platform builds, containerized deployment, legacy modernization
Governance
EU AI Act, auditability & explainability, data residency, privacy architecture, EBA / Basel / FSB
Delivery
end-to-end ownership, PRDs & user journeys, prototype → production, stakeholder alignment

Writing code since age eight — C, C++, Delphi. German (native), English (fluent), French (basic).

Applied through a portal?

Some recruiting portals ask for your CV and then make you type it in again, field by field. I built a small tribute to that experience. It boots like a mainframe and judges you accordingly.

Enter the Null-Device

Next segment

Got a problem everyone has scoped in person-years? Send it over.

matthias.nordwig@programmiere.de

Matthias Nordwig

Technical lead · AI systems in regulated finance

matthias.nordwig@programmiere.de | LinkedIn | Frankfurt Rhine-Main, Germany

Summary

Technical lead with 15+ years in financial services, combining regulated enterprise architecture (SEPA, data warehousing, EBA/Basel) with hands-on building of modern AI systems: audit-ready RAG platforms, agentic workflows and on-device AI in Rust. Turns complex requirements into production systems — and stakeholder gridlock into delivery.

Experience

KfW — Senior Manager (Technical Lead), Data Analytics & AI2018–now, Frankfurt

  • Bootstrapped an internal GenAI R&D function and scaled adoption through enablement and production transition of prototypes.
  • Architected and implemented an audit-ready RAG platform for semantic search across regulatory content, from system design to compliance documentation.
  • Prototyped an autonomous SQL IDE (Codex/DuckDB) that cut complex FSI data integrations to two days.
  • Led end-to-end delivery of enterprise systems with scopes up to 6,000 person-days.

KfW — Senior Manager, Compliance2014–2018, Frankfurt

  • Engineered automated workflows integrating regulatory horizon scanning with internal impact assessments (5,000 person-days).
  • Designed global compliance processes across 100+ international offices.
  • Delivered analytical frameworks for regulatory impact assessments (EBA, Basel, FSB).

KfW — Project Manager, Inhouse Consulting2012–2014, Frankfurt

  • Designed a project and financial steering framework that secured executive approval for a major compliance transformation.
  • Delivered C-level target operating models for reputational risk and KYC.

BearingPoint — Consultant → Manager2007–2012, Frankfurt

  • Directed design and implementation of a credit early-warning system (6,000 person-days), scaled from HVB to the wider UniCredit group.
  • Led data architecture and integration for a SEPA Direct Debit rollout at a major Landesbank (1,500 person-days).
  • Resolved six months of stakeholder gridlock and delivered an FSA liquidity architecture in three months by reverse-engineering a proprietary data mart.
  • Built the Risk Management & Early Warning consulting practice from scratch.

entory AG — Consultant2006–2007, Germany

  • Stepped up as technical lead to rescue a critically escalated data-warehouse project (2,000 person-days) and secured delivery.
  • Engineered a self-service ETL architecture so business units could model shifting IFRS KPIs without blocking engineering.

Selected projects

Skills

Education, publications, languages