AI Engineering · DACH

AI systems that adapt to your needs — not the other way around.

I help companies in the DACH region deploy AI where it truly makes an impact — from use-case identification through architecture to production operations.

Discuss your project →

AI projects rarely fail because of the technology. They fail because the solution doesn't match the problem.

A RAG system optimized for the wrong document type. A chatbot that misses what users actually need. An AI strategy that impresses technically but doesn't improve a single business process.

The root cause is almost always the same: they started with the technology, not with the need.

Background

AI Engineer with roots in cloud infrastructure and solutions architecture in financial services.

I don't just understand the models — I also build and operate the infrastructure they run on. And I know the business processes they're meant to improve.

My Approach

Understand first, then build.

Find the right lever

Not every problem needs an LLM. Not every use case needs self-hosting. And not every task needs an agent. The first question is always: What's the actual bottleneck — and which solution delivers the most value at a reasonable cost?

End-to-end instead of handoff interfaces

I don't take on AI projects in slices — one consultant for strategy, another for implementation, a third for operations. I cover the entire journey: from use-case identification through implementation to production operations.

Transparent and transferable

Everything I build, I document so that internal teams can understand, operate, and evolve it. No black box, no vendor lock-in on me personally. My goal is for the system to work without me — not to need me permanently.

Services

What I can build for you

Every project starts with your needs — not with my tech stack.

AI Strategy & Use-Case Identification

Where can AI actually make a difference in your organization? I identify the use cases with the highest leverage, evaluate the technical options, and deliver a roadmap with concrete next steps — including a PoC or MVP if desired.

RAG Systems — Systematic, Not Standard

Chunking, embedding, retrieval, reranking, synthesis — each layer experimented and optimized in isolation on your data. All experiments tracked transparently, validated not on generic benchmarks but on your documents.

Agentic AI

Tool use, multi-agent coordination, evaluation, and guardrails. From architecture through deployment to operations — on your own infrastructure if required.

MLOps

Automated pipelines, quality gates, model registry, monitoring, rollback. I build the foundation on Kubernetes — so your system isn't just deployed, but truly operated.

Self-Hosted AI

When GDPR, BaFin requirements, or trade secrets rule out external APIs: serving, fine-tuning, embedding models, and evaluation run locally — no API call leaves your network.

Portfolio