AI Systems
& Data Infrastructure

Most organizations sit on data they either process manually or don't process at all. Contracts in PDFs, financial filings in legacy formats, regulatory documents across jurisdictions – valuable information maintained by someone, somewhere, or never touched in the first place.

I build the systems that take over that work. Pipelines, agents, scoring engines – whatever the problem actually needs. Part of that is knowing when a language model solves the problem and when a classical method would have done it faster, cheaper, and better.

The infrastructure depends on the constraints. Data sensitivity, regulation, and cost decide whether that means cloud APIs, local models, or hybrid setups that keep sensitive data where it belongs. The model is usually the easy part. The real work is everything around it.

One thing I keep coming back to: an open-source MCP server over the Austrian company registry – the kind of infrastructure that lets AI agents do real work against real data, not just answer questions about themselves.

Based in Vienna. Background in Economics (WU Wien), International Management (Bocconi), and Information and Communication Systems (FH Technikum Wien, completed while working full-time). I work best in small, technically capable teams with short decision cycles – the kind of setup where things actually ship.

More on what I build and what breaks in the Notes section. If any of this sounds relevant, let's talk.