Max Robotics — Global Service Robot Directory + US Certification

About

Rohan Chitnis is a robotics and AI researcher whose work sits at the intersection of task and motion planning (TAMP), machine learning, and autonomous decision-making. He is perhaps best known for his foundational contribution to combined task and motion planning, with his 2014 paper introducing a planner-independent interface layer garnering an impressive 474 citations — a landmark work that enabled researchers to leverage off-the-shelf planners rather than building costly integrated systems from scratch. His subsequent research expanded this framework into belief space under uncertainty, and later explored how graph neural networks can help robots efficiently navigate planning problems involving hundreds of objects by learning to identify what matters most. A recurring theme across Chitnis's career is making planning tractable and generalizable. His work on Neuro-Symbolic Relational Transition Models (NSRTs) and predicate invention tackles the challenge of learning structured abstractions that enable bilevel planning across continuous state and action spaces. He has also contributed to relational model-based reinforcement learning through the GLIB exploration framework and to real-world object search under sensor uncertainty. Together, his publications reflect a coherent research vision: building robots that can learn compact, interpretable representations of their world and use them to plan intelligently over long horizons.