Max Robotics — Global Service Robot Directory + US Certification

About

Thanard Kurutach is a robotics and artificial intelligence researcher whose work sits at the intersection of visual reinforcement learning, robotic manipulation, and model-based planning. His research focuses on enabling robots to learn complex manipulation tasks directly from visual observations, with a particular emphasis on reducing the need for human demonstrations and overcoming the notorious sample inefficiency of reinforcement learning. Kurutach's most influential contribution, "Learning to Manipulate Deformable Objects without Demonstrations" (2020, 164 citations), introduced a novel iterative pick-place action space combined with model-free visual RL to tackle one of robotics' most challenging open problems — manipulating flexible, non-rigid objects. His complementary work on visual planning and acting (91 citations) demonstrated how robots can reason about object interactions through learned visual models, extending manipulation capabilities beyond rigid-body domains into more realistic domestic and industrial settings. His 2020 paper on Hallucinative Topological Memory further advanced zero-shot visual planning by enabling agents to generalize to unseen goals without additional training. Early work on probabilistic world modeling using Dirichlet-process mixtures reflects his broader interest in robust robot perception. Together, these contributions represent meaningful strides toward sample-efficient, demonstration-free robot learning in complex real-world environments.