About

I came to biology sideways, with an electrical engineering degree first, then a PhD in computational biology at UCSF working on inference problems in single-cell gene expression, followed by a postdoc at the Broad Institute focused on machine learning for massive-scale single-cell data.

Cell Systems, April 2016, cover feature for the shallow-sequencing paper from that PhD work.

That path left me convinced of something that still shapes what I work on: the most useful ideas in this field usually come from treating biological data as a hard information-retrieval and representation-learning problem, and just as much a biological one. Single-cell genomics has produced enormous datasets, but the tools to search, compare, and reason over them at scale are still remarkably underdeveloped.

At Genentech, I lead a research group applying that lens to real drug discovery problems: target discovery, biomarker discovery, and increasingly, patient-level modeling. I think patient-level modeling in particular is one of the places where AI can have near-term transformational impact in biology, in contrast to new-target discovery, which has to survive a decade-long pipeline before it means anything.