My research focuses on designing machine learning algorithms to enable novel and automatic biological discoveries. I also study how to make these discoveries trustworthy and usable to domain scientists, which entails producing mechanistic interpretability and having reliable guarantees.