AI engineer and Software Builder, based in Bonn, Germany. I work on agentic systems and RAG pipelines, the kind of Software integrated with AI that has to keep working long after the demo is over.
About
I'm a Computer Science Master's student at the University of Bonn, focused on applied machine learning: large language models, retrieval, and the engineering that turns them into products people actually use.
At yumo I'm the first technical hire, and I built the core product from the ground up. It takes a structured intake form for an AI teammate and runs it through a three-stage AI pipeline, a completeness check, a quality review, and parallel artifact generation, producing four ready-to-use documents (a Profile, a Recipe, a System Prompt, and User Stories) in under ten minutes. The same work used to take days of manual back-and-forth.
I owned the full build: the Next.js application and multi-step review flow with a hard approval gate, the REST APIs and a custom JWT auth layer wiring it to Supabase, three n8n workflows that run AI agents in parallel and write results straight to the database, and the technical documentation behind all of it.
I care about scoping problems clearly and shipping work that actually gets finished. I read papers, but I'd rather put a working prototype in front of users than a perfect plan on paper.
Projects
Problem. Driving footage carries segmentation, depth, and motion signals — but combining them into a single trainable pipeline is fiddly.
Outcome. Built a multi-task pipeline in CARLA: DeepLabv3-ResNet50 for segmentation, monocular depth estimation, and GRU/ConvGRU heads for ego-motion. Reported IoU and depth metrics with trajectory and point-cloud visualisations.
Contact
The best way to reach me is by email or LinkedIn. I'm always open to new opportunities, collaborations, and good conversations about engineering and building products.