Robotics Simulation Scientist: GPU RL & Sim-to-Real

Amazon; San Francisco (CA); On-site; USD 142,800 - 193,200; Full time

Job summary

Amazon Robotics R&D is seeking an Applied Scientist II to design, build, and validate simulation environments and policy training pipelines that enable robots to learn manipulation and mobility skills in simulation and transfer them to real hardware.

You will work at the intersection of robotics simulation science and modern Physical AI, building GPU-accelerated RL environments, applying domain randomization, and collaborating with SDEs, Technical Artists, and partner science teams.

Qualifications

• PhD or Master’s degree required and robotics research experience.
• Proficiency with ML frameworks (JAX, PyTorch) and robotics simulators (MuJoCo, Isaac Sim).
• 2+ years of experience with physics-based robot learning and sim-to-real transfer.

Responsibilities

• Design and implement GPU-accelerated RL and imitation learning environments in NVIDIA Isaac Lab for manipulation and mobility tasks.
• Build and maintain policy training pipelines for diverse model architectures and evaluate trained policies in simulation.
• Characterize and reduce sim-to-real gaps through systematic validation against real-world data and implement improvements.
• Implement domain randomization strategies to improve policy robustness and transfer to real hardware.
• Develop sim-to-real transfer techniques including system identification and physics parameter calibration.
• Create robot embodiment validation tests to ensure digital twins match real hardware.
• Collaborate with SDEs, Technical Artists, and partner science teams to meet training requirements.

Skills

Reinforcement learning

Imitation learning

PyTorch

MuJoCo

Isaac Sim

URDF

ROS2

Domain randomization

GPU acceleration

Python

Education

PhD or Master’s

Tools

JAX

vLLM

Dynamo

TorchXLA

TensorRT

MuJoCo

Job description

Amazon Robotics R&D is seeking an Applied Scientist II to design, build, and validate simulation environments and policy training pipelines that enable robots to learn manipulation and mobility skills in simulation and transfer them to real hardware.

You will work at the intersection of robotics simulation science and modern Physical AI, building GPU-accelerated RL environments, applying domain randomization, and collaborating with SDEs, Technical Artists, and partner science teams.