Hi, I’m Tyler Vergho

I’m a member of technical staff at Convergent, where we’re working on making AI agents reliable: catching the ways they fail in production and turning those failures into tests they have to pass.

Before Convergent I spent two years building core AI infrastructure at Nooks, and before that I was the first US engineer at Mercor. I studied computer science at Dartmouth and did policy debate, which I still coach.

On the side I built Logos, a search engine for debate evidence.

The best way to reach me is email. I’m also on GitHub, LinkedIn, and X.