Senior Applied Scientist, Safe Legged Locomotion

Amazon; Massachusetts; On-site; USD 167,000 - 226,000; Full time

Benefits offered by this job

Health insurance

RSU eligibility

Paid time off

Job summary

Amazon Robotics, Compass Team seeks an Applied Scientist to develop safe legged locomotion algorithms and deploy them on physical hardware, enabling quadrupeds and humanoids to walk, run, and recover from disturbances with robust, dynamic gaits.

You will leverage RL, sim-to-real transfer, and learning‑based architectures to train policies that interface with model‑based controllers, collaborate with perception, planning and safety teams, and publish results at top robotics venues.

Qualifications

• PhD or Master’s degree with 6+ years of applied research experience.
• 3+ years applying RL to physical robotic systems beyond simulation.
• Proven sim‑to‑real transfer experience for locomotion or manipulation tasks.
• Strong understanding of legged robot dynamics, contact mechanics, and whole‑body control.
• Proficient in Python and deep learning frameworks with custom RL training pipelines.
• Experience with physics simulators such as Isaac Gym, MuJoCo or PyBullet; publications at top venues.

Responsibilities

• Design, train, and deploy RL policies for legged locomotion on quadruped and humanoid platforms.
• Collaborate with safety teams to enforce safety bounds using constraints during or after training.
• Develop sim‑to‑real transfer pipelines to bridge reality gap with domain randomization and system ID.
• Integrate learned policies with model‑based whole‑body controllers and optimization layers.
• Create reward functions and curricula balancing performance and safety constraints.
• Build large‑scale training infrastructure including simulators and parallel pipelines.
• Evaluate policy performance through simulation, hardware experiments, and failure analysis.
• Explore emerging techniques and publish results at top robotics and ML venues.

Skills

RL for robotics

Python

Deep learning

Robotics dynamics

Publications

Education

PhD or Master’s+6 yrs

Tools

Isaac Gym

MuJoCo

PyBullet

PyTorch

JAX

Job description

Amazon Robotics, Compass Team seeks an Applied Scientist to develop safe legged locomotion algorithms and deploy them on physical hardware, enabling quadrupeds and humanoids to walk, run, and recover from disturbances with robust, dynamic gaits.

You will leverage RL, sim-to-real transfer, and learning‑based architectures to train policies that interface with model‑based controllers, collaborate with perception, planning and safety teams, and publish results at top robotics venues.