The Medvolt Small Molecule NCE Platform leverages advanced AI and computational techniques to design new molecules with desired pharmacological properties. It uses fragment-based modeling, generative AI, and reinforcement learning to create and optimize drug candidates.
The platform also employs molecular docking and multi-task graph neural networks for virtual screening, assessing drug-likeness, and predicting ADME, physicochemical, and toxicity properties. State-of-the-art molecular dynamics simulations and alchemical free energy calculations refine the process for accurate protein-ligand binding predictions. Technologies used include AWS Elastic Beanstalk, Docker, RDS, S3, AWS Batch, FastAPI, and DocumentDB.