The Koo Lab at the Simons Center for Quantitative Biology at Cold Spring Harbor Laboratory develops machine learning methods to study gene regulation and biological sequence function, with an emphasis on interpretability, generalization, and mechanistic insight. Our work focuses on building and evaluating deep learning models that link DNA and protein sequence to molecular and cellular phenotypes, developing principled methods to interpret what these models learn, and integrating modeling with experimental perturbation data. We are particularly interested in understanding how regulatory logic is encoded in sequence, how model predictions vary across genetic and cellular contexts, and how AI-based models can be audited and refined to support biological discovery. Through these efforts, our goal is to move beyond black-box prediction toward computational frameworks that yield testable hypotheses and mechanistic understanding in regulatory genomics and cancer biology.