Max Robotics — Global Service Robot Directory + US Certification

About

Kuan Fang is a robotics and machine learning researcher whose work sits at the intersection of robotic manipulation, embodied AI, and learning from demonstration. His research addresses some of the field's most pressing challenges: enabling robots to operate intelligently in complex, real-world environments with limited supervision and partial observability. Fang has made significant contributions to task-oriented grasping and tool manipulation, demonstrating through his widely cited work — including "Learning Task-Oriented Grasping for Tool Manipulation from Simulated Self-Supervision" (185 citations) — that robots can learn meaningful, goal-aware grasping strategies through simulated self-supervision rather than costly real-world data collection. His "Scene Memory Transformer" (186 citations) introduced a powerful approach for equipping embodied agents with long-term memory in partially observable, long-horizon tasks. Additional notable contributions include "Demo2Vec" (108 citations), which leverages online video demonstrations to reason about object affordances, and work on 6-DoF grasp detection via implicit representations (114 citations), highlighting the synergy between 3D reconstruction and grasp learning. Across his portfolio, Fang consistently advances scalable, data-efficient frameworks that bridge simulation and real-world deployment — making his research highly relevant to researchers working on autonomous robotics, human-robot interaction, and generalizable manipulation systems.