Max Robotics — Global Service Robot Directory + US Certification

About

Andrew Szot is a leading researcher in embodied AI, robotics, and simulation, whose work is redefining how virtual and physical agents interact with complex, human-centric environments. His most impactful contributions center on developing high-fidelity simulation platforms and scalable learning frameworks for mobile manipulation and human-robot collaboration. Szot is the lead author of Habitat 2.0 (25 citations), a groundbreaking simulation platform that trains virtual robots in interactive 3D spaces with complex physics, and Habitat 3.0 (15 citations), which extends this to co-habitation with human avatars for collaborative tasks. His Skill Transformer (13 citations) introduces a monolithic policy that combines conditional sequence modeling with skill modularity to solve long-horizon robotic tasks, while his work on Adaptive Coordination (5 citations) pioneers multi-agent social rearrangement in simulated home environments. Additionally, Galactic (4 citations) achieves remarkable scalability in end-to-end reinforcement learning for rearrangement tasks at 100,000 steps per second. Through these innovations, Szot has established himself as a key architect of the next generation of embodied AI systems, enabling robots to learn, adapt, and collaborate in the messy, dynamic spaces of everyday life.