Adaptive Agentic Robotic Systems

This project focuses on creating robotic systems that can adapt and improve during deployment in dynamic, unstructured environments such as warehouses and industrial sites. It combines the reliability of classical robotics with the flexibility of learned policies, enabling robots to reason, plan, and recover from errors. The research explores reinforcement learning for long-term orchestration improvements and in-context learning for rapid, test-time adaptation using rich feedback like language or demonstrations. These innovations aim to deliver robots that are more reliable, generalizable, and responsive, paving the way for safe, intuitive automation in logistics, manufacturing, healthcare, and home settings.