This project explores the development of privacy-preserving, multi-agent AI systems that use large language models and differential privacy. The goal is to design decentralized agent frameworks that can collaborate and reason collectively while maintaining strict data protection. We intend to develop a unique system aimed to support complex decision-making tasks across sensitive domains, building toward safer, fairer, and more transparent AI applications.

I am doing SuperUROP because I want to continue my research in privacy and ethical AI at a deeper level by pursuing a more involved project with my lab. I want to be able to present interesting and impactful work and publish something meaningful to my community.