Agility Robotics is looking to develop and deploy reinforcement learning models for whole-body robot control, integrating perception to enable locally collision-free locomotion and manipulation in real-world environments for their robot, Digit.
Requirements
- 3+ years of experience developing and deploying reinforcement learning models for robotics applications.
- Strong programming skills in Python, with proficiency in deep learning frameworks such as PyTorch.
- Experience designing reward functions, tuning hyperparameters, and implementing exploration strategies to solve complex control tasks.
- Proven experience with perception-in-the-loop control, integrating real-time sensory inputs for reactive or adaptive behaviors.
- Familiarity with robot simulation environments (e.g. Mujoco, Isaac Sim) and sim-to-real transfer techniques.
- Experience with C++ for integration of learned controllers into real-time control systems.
Responsibilities
- Design, train, and deploy robust reinforcement learning policies for locomotion, manipulation, and dynamic interactions with the environment.
- Integrate perception inputs into control policies to achieve obstacle-aware, collision-free motion.
- Develop and maintain core reinforcement learning infrastructure, including scalable training pipelines and evaluation frameworks.
- Design and implement new simulation environments and tasks to support training and deployment of control policies.
- Collaborate with robot software and deployment teams to ship production-quality policies to Digit.
Other
- Ability to work collaboratively in a fast-paced environment to deliver safe, high-quality software.
- Advanced degree (MS or PhD) in Robotics, Computer Science, or a related field.
- Experience deploying reinforcement learning policies on real-world bipedal or quadrupedal robots.
- Publications in top ML or robotics conferences (e.g. NeurIPS, ICML, CoRL, RSS, ICRA).
