Max Robotics — Global Service Robot Directory + US Certification

About

Eric Undersander is a robotics and embodied AI researcher whose work sits at the intersection of robot learning, simulation, and mobile manipulation. He has made significant contributions to the development of simulation platforms for training home assistant robots, most notably through the Habitat series — including Habitat 2.0 and Habitat 3.0 — which provide rich, physics-enabled environments for studying complex rearrangement and collaborative human-robot tasks. His research on Habitat-Web (75 citations) demonstrated how large-scale imitation of human demonstrations can teach robots effective object-search strategies, while his work on Adaptive Skill Coordination (ASC) advances long-horizon mobile manipulation by combining modular visuomotor skills with adaptive coordination mechanisms. Beyond navigation, Undersander has explored graph neural networks for interpretable robot manipulation (44 citations), framing task learning as structured relational reasoning over object graphs. His Galactic framework pushes the boundaries of reinforcement learning scalability, achieving 100,000 simulation steps per second for rearrangement training. Together, his publications reflect a coherent vision: building robots that can efficiently, intelligently, and adaptably operate in real human environments — a goal increasingly relevant as embodied AI moves toward practical deployment.