BrassBunny Alchemist for Drug Safety

Our Project: BrassBunny Alchemist for Drug Safety Inspiration The need for rapid drug development to serve an aging population and improve overall quality of life is a significant driver in healthcare. However, a major hurdle is the prevalence of adverse drug events, with drug-drug interactions (DDIs) being a primary cause. We were inspired to create a solution that could proactively address this challenge, making drug development safer and more efficient. How We Built It BrassBunny Alchemist is an AI-powered system that leverages intelligent agent collaboration. We used Temporal on AWS to orchestrate these agents, ensuring a scalable and robust architecture. Our system employs advanced machine learning models, along with Claude 3.7 Sonnet, to predict potential DDIs and suggest safer alternative drug combinations. The Message Passing and Coordination Pattern (MCP/A2A) is integral to our design, enabling seamless communication and coordination between the agents. Challenges We Ran Into Developing a system to accurately model the complex mechanisms of DDIs presented a significant challenge. DDIs are influenced by a multitude of factors, and capturing these nuances in a computational model is difficult. Additionally, integrating data from diverse sources, each with its own format and quality, required substantial effort. What We Learned This project reinforced the power of agent-based systems in tackling complex biomedical problems. We learned that by breaking down the problem into smaller, manageable tasks and assigning them to specialized agents, we could achieve a more robust and scalable solution. Furthermore, we gained a deeper appreciation for the importance of robust workflow orchestration in managing the interactions between these agents.

Built With

• amazon-web-services
• claude
• rime
• temporal