Senior AI Engineer building reliable production GenAI systems, evaluation platforms, and agentic infrastructure.

I build the evaluation, guardrail, and observability capabilities required to operate GenAI systems reliably in production.

My work spans RAG quality, agentic workflows, LLM evaluation, tracing, and the platform architecture needed to make AI systems measurable, debuggable, and safe.

See AgentGuard for a concrete example of this work.

AgentGuard — Production-grade QA and reliability layer for agentic AI, covering guardrails, RAG evaluation, red teaming, PII detection, and tracing for safer LLM applications.

JobWatch-CH — End-to-end AI product built with TypeScript, React, Firebase, and GCP that ingests Swiss job listings in real time and uses Gemini to generate structured insights and matching signals.

Dark Software Factory — Experimental multi-agent software delivery pipeline exploring specialist agents, orchestration patterns, and feedback loops for autonomous engineering workflows.

Homelab MLOps Stack — Local platform for model serving, fine-tuning, and LLM experimentation, integrated with GCP and Vertex AI.

Languages & Backend: Python, FastAPI, TypeScript
Cloud & Platform: Docker, GCP, Vertex AI
LLMOps & Agentic AI: LangChain, LangGraph, Qdrant, Langfuse, DeepEval, OpenTelemetry, MCP

Staff / Principal AI Engineer roles focused on GenAI platforms, AI reliability, and production LLM systems in the DACH region.
Open to hybrid or remote opportunities.
French (native), English (fluent), German (C1).

When not building AI systems, I'm on a mountain bike.