Makoto Sato

Makoto Sato is a Ph.D. student in the Matsuo-Iwasawa Laboratory at The University of Tokyo, where he researches robot learning with a focus on scalable imitation learning and test-time scaling for robot foundation models. He has collaborated with Sakana AI, AIRoA, the Matsuo Institute, and AIST, and has taken on research internships at NVIDIA and Google DeepMind. His published work includes task and motion planning using residual reinforcement learning for long-horizon manipulation and test-time scaling for in-context imitation learning with vision-language models.