Adam Gayoso

I am a Senior Research Scientist at Google DeepMind, where I work on genomics. My projects at GDM include AlphaGenome, a model that predicts how genetic variants affect regulatory activity directly from DNA sequence.

Previously, I completed my PhD in the Center for Computational Biology at UC Berkeley, co-advised by Aaron Streets and Nir Yosef. There I developed scvi-tools, a Python library for probabilistic analysis of single-cell omics data, with implementations of published models and the building blocks for new ones. Models I designed with it include totalVI, for multimodal single-cell integration, and veloVI, for transcriptional dynamics. I also co-founded the scverse consortium.

Before that, I studied operations research and computer science at Columbia University, where I was introduced to single-cell genomics in the Dana Pe’er Lab.

Selected work

• AlphaGenome — a unified DNA sequence model predicting how genetic variants affect regulatory activity across the genome. Nature, 2026. code docs
• scvi-tools — probabilistic models for single-cell omics, and a framework for building new ones. Nature Biotechnology, 2022. code docs
• Publications on Google Scholar.