Dominik Tilman

I build systems that pull the signal out of noisy data, so you can act and rely on it.

Approach

Data is cheap. Knowing what to trust isn't.

Messy inputs, high stakes, a call that has to be right. That's my favorite kind of problem. Hard facts, subjective opinion, uncertain forecasts: each is weighted by how well it's actually held up. So you don't just get an answer. You get how far to trust it: where it's solid, where it's thin, how much weight it can bear. The reasoning stays in the open, not a model you have to take on faith. It sharpens over time, as the models learn what holds up and what doesn't.