Senior Machine Learning Engineer, RL / Locomotion

ABOUT THE TEAM

The Anduril Frontier Systems team is building the next generation of robotic platforms for defense and industrial applications. We are a small, high-performing team of roboticists, ML engineers, and systems engineers delivering real-world capability to operators in the field. Our systems operate in unstructured, contested environments where robustness and reliability are non-negotiable. We seek to deliver validated, integrated capabilities to the Department of Defense and the Intelligence Community.

ABOUT THE JOB

We are seeking a Senior Machine Learning Engineer who specializes in RL / Locomotion to develop and deploy locomotion policies for legged robotic platforms. You will own the full pipeline from simulation training to real-world deployment, building systems that enable robust mobility across challenging terrain - rubble, stairs, slopes, and degraded environments. Your work will directly determine whether our platforms can operate where warfighters need them.

WHAT YOU’LL DO

• Design, train, and deploy reinforcement learning policies for legged robot locomotion using GPU-parallelized simulation (Isaac Gym / Isaac Lab)
• Develop terrain curriculum and domain randomization strategies that produce policies robust to real-world conditions
• Own the sim-to-real transfer pipeline, identifying and closing reality gaps
• Train policies for stair climbing, rough terrain traversal, payload carry, push recovery, and fall recovery
• Define and evaluate performance metrics for locomotion robustness
• Collaborate with manipulation and perception engineers to integrate locomotion into a full autonomy stack

REQUIRED QUALIFICATIONS

• 3 - 8+ years of experience with reinforcement learning for legged or mobile robots
• Strong background in dynamics, control, and robot locomotion
• Experience with RL algorithms (PPO, SAC) and training frameworks (rsl_rl, Stable Baselines, rl_games)
• Hands-on experience with physics simulation (Isaac Gym, MuJoCo, PyBullet)
• Demonstrated sim-to-real transfer on physical robotic systems
• Proficiency in Python and PyTorch
• Eligible to obtain and maintain a U.S. security clearance

PREFERRED QUALIFICATIONS

• Experience with bipedal or multi-limbed robotic platforms
• Experience with NVIDIA Isaac Lab or Omniverse
• Publications in top robotics venues (RSS, CoRL, ICRA, IROS)
• Prior work in defense technology or startups