San Francisco · Pre-company

Leah Matzat

Scientist and builder. A decade scaling clinical genomics at Invitae. Now focused on what it takes to train biological foundation models on data that's actually trustworthy.

The AI Virtual Cell is the most important bet in biology right now. But the models are only as good as the data they train on — and that problem is largely unsolved.

I know what clinical-grade biological data looks like when it's built right. I'm pre-company, deliberately. Talking to researchers, builders, and investors to pressure-test the thesis before I lock in.