Based in Zurich. I lead engineering for product companies — hands-on from architecture to delivery — with emphasis on clear UX and pragmatic AI. Serial CTO; two projects acquired. Recent work includes applied LLMs and Gen-AI products. Now building in the interection of AI and legal tech.
I'm a builder at heart: I like turning messy problems into working software — especially where product, UX, and backend meet. Emerging tech is part of the job (AI, Web3 where it made sense); the constant is shipping something people can actually use.
What motivates the career side is simple: help teams build companies and products that outlast the hype cycle — solid engineering, honest scope, measurable outcomes.
Small products and experiments — clear UX, privacy-minded defaults, something measurable at the end.
German vocabulary with a daily, measurable workflow.
What it is
Flashcards for ~850 high-frequency words: progress, reviews, accuracy, streaks. Progress stays in the browser (IndexedDB).
Why I built it
Language learning fails when progress is invisible or the word list is wrong. This keeps the loop tight and local-first.
Speech-to-text in the browser, after a one-time model load.
What it is
Real-time transcription with Whisper Tiny on-device, quick updates, export for use elsewhere.
Why I built it
Voice tools do not have to send audio to a server. This stays in-session and still feels fast.
How to build and keep a motivated engineering team
A step-by-step guide to creating a llms.txt file in Next.js
A step-by-step guide to configuring Bun debug in VSCode
