Max Robotics — Global Service Robot Directory + US Certification

About

Gabriel B. Margolis is a robotics researcher specializing in reinforcement learning for legged locomotion and agile robot control. His work sits at the intersection of machine learning and real-world robotics, with a particular focus on enabling quadruped robots to move with speed, adaptability, and dexterity in unstructured environments. Margolis is perhaps best known for his landmark research on rapid locomotion via reinforcement learning, which produced an end-to-end learned controller that set a record for the MIT Mini Cheetah, achieving sustained speeds of up to 3.9 m/s on challenging natural terrains. This work has accumulated over 220 citations across multiple publication venues, underscoring its significant influence on the field. His research on DribbleBot extended these principles into dynamic manipulation, demonstrating that a legged robot could dribble a soccer ball under real-world conditions — a compelling showcase of agility meeting dexterity. Through projects like "Walk These Ways," Margolis has also tackled the challenge of generalizable locomotion policies, introducing frameworks that allow robots to adapt their behavior across out-of-distribution environments without costly retraining. His contributions are shaping a new generation of robots capable of operating effectively beyond the laboratory.