Max Robotics — Global Service Robot Directory + US Certification

About

Federico Casarini is a robotics researcher pushing the boundaries of what legged robots can achieve by combining deep reinforcement learning with insights from biology. His work centers on synthesizing agile, reusable movement skills for both bipedal and quadruped robots, enabling them to perform complex behaviors in dynamic environments. His most impactful contribution, "Learning agile soccer skills for a bipedal robot with deep reinforcement learning" (2024, 147 citations), demonstrates that deep RL can train a low-cost humanoid to play simplified one-versus-one soccer, composing sophisticated skills like dribbling and shooting. Casarini also explores skill transfer through imitation in "Imitate and Repurpose" (2022, 20 citations), where prior knowledge from human and animal motion capture data is repurposed for real robots. His work on "Barkour" (2023, 13 citations) establishes benchmarks for animal-level agility in quadrupeds, while "Beyond Pick-and-Place" (2021, 16 citations) tackles the challenge of stacking objects with complex geometry. Notably, his 2025 work on "Gemini Robotics" (4 citations) signals a move toward integrating large multimodal models into physical agents. With over 200 total citations, Casarini is a rising figure in learning-based robotics, bridging simulation and real-world deployment.