How I work

My work is mostly decision logic — the rules, models and arithmetic a system uses to choose what to do — and the engineering that carries it. As co-founder and CTO of a young company, I've built its platform and taken it to production, run automation against live systems where mistakes cost real money, and built the monitoring that shows whether any of it is behaving.

The thread is that a method only counts once it runs. An algorithm that works in a notebook and one that survives real data, real costs and real failure modes are not the same object — most of the difficulty lives in the distance between them, and that distance is the part I like.

I also like the writing half of engineering. A co-authored, peer-reviewed paper and national research funding from Austria's FFG both forced the same discipline: state the problem precisely, defend the approach. An architecture that survives that is usually a better one.

None of this is bound to one domain. Modeling a messy problem, hardening an algorithm for production, keeping a system observable — these carry over to anywhere decisions are automated.