About
About
I'm Gerhard, a software architect and engineer. Designing and building software is what I enjoy most — it has been for twenty years, and it still is.
I like to understand the details without losing sight of the bigger picture. I began with embedded software: realtime C++ for optical inspection, then drivers and protocols for measurement devices. There I got interested in the web browser as a user interface, and spent several years bringing web technology to embedded hardware. Along the way I became a polyglot: C and C++, TypeScript, Python, Lua, Java and GLSL — I pick the language that fits the problem, not the other way round. Later I led a frontend team for a job search engine, built a WebGL seat map for a ticketing platform, and today I work on AI at the edge.
Across these roles I've focused on efficient, secure and maintainable systems — with good tests, observability and reliable CI/CD. AI is part of how I work today; knowing the code well helps me decide what to delegate and how to judge the result. When leading teams, I prefer to stay close to the code, support people through reviews and keep priorities clear.
Now · 2026
- Building
- A narrow-focused LLM inference engine for Apple Silicon and other UMA platforms
- Exploring
- Agents running on local inference
- Following
- NPUs, accelerators and on-device models
Facts
- Working
- Remote & Rhein-Main area
- Trained as
- Dipl.-Ing. Photonics & Image Processing
- Working since
- 2005
- Studied at
- Hochschule Darmstadt