Max Robotics — Global Service Robot Directory + US Certification

About

Zuxin Liu is a robotics and artificial intelligence researcher whose work spans autonomous navigation, multi-agent systems, and safe reinforcement learning. He first gained widespread recognition with DS-SLAM (2018), a pioneering semantic visual SLAM system designed to handle dynamic real-world environments — a long-standing challenge in mobile robotics — which has accumulated over 1,050 citations and established him as a significant voice in the SLAM community. Building on his robotics foundations, Liu extended his research to multi-agent coordination, developing MAPPER, a decentralized evolutionary reinforcement learning framework enabling large fleets of robots to navigate complex, unpredictable environments, garnering over 100 additional citations. In more recent years, Liu has pivoted toward the critical challenge of safe reinforcement learning, addressing how AI agents can satisfy safety constraints before deployment in high-stakes applications. His constrained variational policy optimization framework tackles the instability and optimality shortcomings of prior methods, while subsequent contributions include benchmark datasets for offline safe RL and methods for learning shared safety constraints from demonstrations. This trajectory reflects a researcher steadily bridging cutting-edge robotics with principled safety guarantees — work of growing importance as autonomous systems move from controlled settings into real-world deployment.