Max Robotics — Global Service Robot Directory + US Certification

About

Souradip Chakraborty is a researcher at the forefront of reinforcement learning (RL) and robotics, with a focus on overcoming fundamental challenges in continuous control and navigation. His work addresses the critical problem of sparse rewards—a scenario where feedback is infrequent, making learning inefficient. To tackle this, Chakraborty introduced the Heavy-Tailed Stochastic Policy Gradient (HT-PSG) algorithm, which leverages heavy-tailed distributions to explore more effectively in sparse-reward environments, a contribution that has garnered early recognition with 3 citations. He further extended this approach to outdoor robot navigation with the HTRON system, demonstrating practical improvements in real-world deployment. Chakraborty also explores adaptive policy design through language-based feedback (RE-MOVE), enabling robots to request human assistance in dynamic environments. Most notably, his 2024 paper on the vulnerability of LLM/VLM-controlled robotics highlights critical security and reliability risks in integrating large language models with robotic systems, a timely and impactful contribution as these technologies become more prevalent. With a growing citation record and a focus on bridging theoretical RL advances with robust, real-world robotic systems, Chakraborty is establishing himself as a key voice in safe and adaptive autonomous navigation.