Not a straight line.

I'm Niklaas. I studied political science in Montana and somehow ended up building AI systems for go-to-market teams in Germany. Each step made sense at the time, even if the whole path doesn't look like a plan.

I got into performance marketing because it was the one corner of the field you could measure. That pulled me into data, then into writing code to make the data trustworthy, then into building agents that do the work on their own. One instinct hasn't changed: I don't take a reported number on faith, and I'd rather build the system that produces an honest one than argue over a dashboard.

The political-science background isn't a detour I apologise for. It's mostly the study of how people, incentives, and systems behave versus how they're supposed to. That turns out to be half the job when you're automating how a team works.

I like the unglamorous parts. The plumbing, the eval loops, the things that have to keep working at 3am when nobody's watching. I care more about a system someone can run without me than a clever demo that falls over in week two. A working thing I can hand off beats a deck about one, every time.

Right now I'm looking for the next hard problem to build for. If any of this sounds like your kind of thing, say hello. The full history and toolkit live in the CV.