We adapt predictive AI/ML to suggest how likely a given set of traits might be present (“a phenotype”) to improve how we might build safer clinical systems and better understand the underlying risks of disease. With an international team of collaborators, we bring scalable AI to precision medicine and postmarket safety surveillance.

Selected Publications:

Scalable incident detection via natural language processing and probabilistic language models

Development and multi-site external validation of a generalizable risk prediction model for bipolar disorder

Significant shared heritability underlies suicide attempt and clinically predicted probability of attempting suicide