Max Robotics — Global Service Robot Directory + US Certification

About

Alexey Dosovitskiy is a prominent researcher at the intersection of robotics, machine learning, and autonomous systems, with particular expertise in reinforcement learning, imitation learning, and agile robot control. His work spans two major domains: legged robotics and autonomous driving, where he has made foundational contributions to how machines learn from experience rather than hand-crafted rules. His most influential work on learning agile motor skills for legged robots (2019, nearly 1,400 citations) demonstrated that reinforcement learning could produce dynamic, animal-like locomotion previously thought unachievable through traditional engineering approaches. In autonomous driving, his research on end-to-end conditional imitation learning (over 1,000 citations) tackled the critical challenge of making learned driving policies responsive to navigational commands at test time — a key step toward practical deployment. Dosovitskiy has also advanced sim-to-real transfer through modularity and abstraction, addressed terrain prediction for legged robots via self-supervised learning, and explored agile drone flight in dynamic environments. His recurring theme is bridging the gap between simulation and real-world performance. With thousands of citations across multiple subfields, his work has significantly shaped modern robot learning research.