I push tools past what they were built for.
That's where the real limits are. Usually the simplest thing that survives is the answer.
Right now: phishing detection at Cisco, Kremlin talking points in Czech comment sections, a browser CLI that lets AI agents drive a real browser, and a camera on servos that watches back.
I'm looking for the problem worth going all in on.
Prague’s best builders should know each other.
I recently joined a hackathon in Berlin and remembered that the quickest way to know someone is to build with them. I want to put 20 people who feel the same in one room in Prague.
selected work
notes
work
about
Theoretical biophysics by training. Open quantum systems, Redfield equations. The math still holds, the abstraction is what comes in handy.
From there: data scientist, then ML engineering. Now focusing on detections.
Off-keyboard I'm on rock: bouldering, sport routes, the odd alpine multipitch.