I developed an end-to-end virtual screening pipeline for HIV-1 protease inhibitors using DeepChem. This workflow integrated ligand-based feature extraction (Morgan fingerprints, RDKit descriptors), molecular docking using AutoDock Vina, and supervised machine learning models (random forests, graph neural networks) to rank candidate molecules. The pipeline automated data preprocessing, model training, docking simulation, and scoring—enabling rapid prioritization of small-molecule inhibitors. The approach was tailored for high-throughput screening and contributed to TCS Life Sciences’ internal toolkit for antiviral compound discovery.