I build ML‑enabled discovery pipelines that combine graph models for molecular properties and compound–protein interactions with explainable AI and high‑throughput screening. The goal is prioritized libraries, fewer assay artifacts, and faster, more autonomous design–make–test cycles.