David Koes focuses on structure-based drug discovery and computational biophysics: molecular docking, pharmacophore modeling, protein-ligand scoring with deep learning, and interactive visualization/search of chemical space via open-source platforms, generative models, and data-driven property prediction.