Amazon Web Services, or AWS, said on April 16 it has launched Amazon Bio Discovery, an AI-based application designed to help scientists design and validate new drugs faster and more systematically.

The company said Amazon Bio Discovery gives scientists direct access to a catalogue of Biological Foundation Models, or bioFMs, specialised AI models trained on large biological datasets. The models generate and evaluate drug candidates and help accelerate antibody therapeutics research in the early stage of drug discovery.

Using Amazon Bio Discovery AI agents, scientists can use natural language to select models suited to research goals, optimise input variables and assess candidates. They can further train models on existing experimental data to improve predictive accuracy. Candidate materials are sent to physical laboratories for synthesis and testing, and results are fed back into the application, creating a lab-in-the-loop structure that enables rapid iterative experiments.

Rajiv Chopra (라지브 초프라), vice president for healthcare AI and life sciences at AWS, said, "AI agents enable researchers without computing expertise to use advanced scientific capabilities." He said, "AI systems help scientists design drug molecules, orchestrate tests, learn from results and become more sophisticated as experiments are repeated." He added, "This combination of advanced AI and the robust, secure infrastructure that AWS has built for regulated industries can accelerate antibody discovery in ways that were previously impossible."