I'm an AI Engineer who turns large language models into things that actually ship — RAG pipelines, agentic orchestration, and multimodal systems that survive contact with production.
I build production-ready, AI-powered systems that bridge large language models with real business workflows — and I genuinely enjoy the messy gap between a slick prototype and something a team can rely on.
Backed by two Anthropic certifications and 3+ years of full-stack engineering, I specialize in RAG pipelines, agentic orchestration, and multimodal architecture. I've taken AI systems from idea to production for research institutions, startups, and SMEs — currently as a Research Assistant at the Leibniz Institute (DIPF) while finishing my Master's in Fulda.
An AI presentation coach that watches you rehearse and gives real-time multimodal feedback — voice, posture, pacing — with heavy video work offloaded to an async Celery/Redis pipeline.
Visit Site →A multi-tenant, RAG-based proposal generator for SMEs. A two-pass LLM pipeline — an Analyst that reads the brief and a Writer that drafts — turns scattered context into polished, on-brand proposals.
Visit Site →A generative-AI music platform built on ambient and systemic music concepts — endless, evolving soundscapes that compose themselves rather than loop. The fun, weird side project that keeps me curious.
An open-source Chrome extension acting as a social media "co-pilot" directly in your browser sidebar. Leveraging the Claude API, it provides real-time, context-aware assistance to draft comments and posts 10x faster on LinkedIn and X.
Visit Site →Deterministic record/replay and regression-testing harness for LangGraph agents. Record every LLM call, tool response, clock value & RNG draw — then replay the exact sequence offline at ~$0 API cost to reproduce failures and diff prompts.
Visit Site →I'm open to remote, on-site, and hybrid roles across Germany — and always up for a good problem. Let's build something.
