Safe RL Robotics Scientist: Dynamic Legged Locomotion

Amazon; Pasadena (CA); On-site; USD 143,000 - 193,000; Full time

Job summary

Amazon is seeking an Applied Scientist to join the SAF Lab to advance safe reinforcement learning for dynamic legged robots. You will design RL policies, integrate with physics models, and ensure safety constraints are learned and enforced from training to run-time across perception, planning, and control components.

The role involves building scalable training infrastructure, publishing in top venues, and collaborating with teams across ML, controls, and mechanical engineering to deploy robust

Qualifications

• PhD in Computer Science, Robotics, Mechanical Engineering, Electrical Engineering, or related field with focus on reinforcement learning, robot learning, or control.
• Experience applying RL to physical robotic systems with sim-to-real transfer on dynamically stable robots.

Responsibilities

• Collaborate with product teams and science leaders to set a science roadmap with impact on real robots.
• Design, train, and deploy RL policies for legged locomotion on physical robots.
• Develop sim-to-real transfer pipelines robust to the reality gap including domain randomization and system identification.
• Integrate control-based methods with RL and safety layers in training and run-time.
• Develop and maintain large-scale training infrastructure for locomotion policy learning.

Skills

Reinforcement learning

Python

PyTorch

JAX

Robotics

Education

PhD in Computer Science/Robotics/Engineering

Tools

Isaac Gym/Sim

MuJoCo

PyBullet

Job description

Amazon is seeking an Applied Scientist to join the SAF Lab to advance safe reinforcement learning for dynamic legged robots. You will design RL policies, integrate with physics models, and ensure safety constraints are learned and enforced from training to run-time across perception, planning, and control components.

The role involves building scalable training infrastructure, publishing in top venues, and collaborating with teams across ML, controls, and mechanical engineering to deploy robust