AI Systems Engineer & AI Infrastructure Engineer
Building AI solutions that improve decisions, automate workflows, and create measurable production impact
Munich · Germany
I help teams identify where AI creates the most value, then ship production solutions that improve speed, quality, and reliability. Across 20+ successful projects, I have delivered pragmatic AI systems with strong customer-facing execution in the German enterprise market.
I focus first on use cases: where AI can reduce manual effort, improve decisions, accelerate delivery, or unlock new product capabilities. I then translate those goals into production AI solutions with clear ownership and measurable outcomes.
I work hands-on from prototype to rollout, combining AI systems engineering with practical implementation: AI platform engineering, model evaluation, CI/CD, observability, and reliable runtime operations.
I also bring experienced consulting leadership: managing customer relationships, stakeholder alignment, and enterprise rollouts in complex German market environments. I am especially interested in production AI, agent engineering, and MCP-enabled systems.
RAG pipelines, prompt engineering, model evaluation, and production delivery for LLM-based systems.
Multi-agent orchestration, tool calling, MCP servers, and robust runtime systems that run reliably.
End-to-end infrastructure for data, model serving, evaluation, and developer workflows.
Production-first delivery with CI/CD, testing, release discipline, and boring-in-operation reliability.
Distributed systems design, Kubernetes operations, observability, and tracing across cloud-native runtimes.
Developer infrastructure, coding agent integration, guardrails, and workflows that improve engineering velocity.
Examples of how I apply AI to solve concrete business and operational problems, backed by hands-on experience building data platforms, ML serving systems, anomaly detection pipelines, and production data quality frameworks.
Addressed the challenge of reviewing large corpora of technical documentation through a production-ready document intelligence workflow with RAG and agentic tool use. I refined the architecture, introduced modern development workflows, and strengthened operations and monitoring, resulting in measurable adoption and efficiency gains.
Made it simple for teams to create internal chat assistants from their own document corpora, reducing setup effort and improving knowledge access. I designed the reusable architecture and turned the first prototype into a production-ready baseline, enabling broad adoption with strong retrieval quality and guardrails.
Targeted better quality and speed for AI-generated code, especially in large and complex codebases, by integrating coding agents into daily engineering workflows. I created coding guidelines and agent instructions, built a specialized review agent, and added internal documentation as context, improving review consistency and development flow.
I run delivery as a practical sequence: ideation workshops, focused PoC, friendly user evaluation, hardening, rollout, and scaling. The goal is always measurable production value: increased employee productivity, better resource allocation, and business solutions that run reliably over time. In implementation, I work hands-on with async Python services on Azure, CI/CD automation, RAG pipelines, and agent workflows with tool calling. Reliability is built in from the start through evaluation gates before rollout, observability and tracing, guardrails and policy checks, human-in-the-loop fallback paths, and continuous cost/performance optimization. I manage customer and stakeholder alignment through close listening, direct communication, and disciplined focus on end-user requirements.
A selection of tools and experiments I have built and open-sourced. Most of these projects are prototypes for vibe-coding and agentic engineering approaches.