Max Robotics — Global Service Robot Directory + US Certification

About

Clayton Tan is a leading researcher at the intersection of robotics, natural language processing, and large-scale machine learning. His work focuses on grounding language in robotic affordances and scaling robot learning through innovative data generation and model architectures. Tan’s most influential contribution is the landmark paper “Do As I Can, Not As I Say” (2022, 516 citations), which introduced a method for leveraging large language models to generate actionable, context-aware commands for robots, bridging the gap between semantic knowledge and physical execution. He also co-developed RT-1: Robotics Transformer for Real-World Control at Scale (2023, 512 citations), a foundational model that demonstrated how transfer learning from diverse, task-agnostic datasets can enable robots to perform a wide range of manipulation tasks with minimal fine-tuning. More recently, Tan has advanced offline reinforcement learning with Q-Transformer (2023, 16 citations), which uses autoregressive Q-functions to train multi-task policies from large offline datasets. His work on semantically imagined experience (2023, 66 citations) further pushes the boundaries of data efficiency by generating synthetic training data. With over 1,100 total citations, Tan is shaping the future of generalist robots capable of understanding and acting on natural language instructions in real-world environments.