Sim-to-Real Robotics ML Engineer

Noble Machines, Inc; Sunnyvale (CA); On-site; USD 170,000 - 400,000; Full time

Job summary

Noble Machines in Sunnyvale, CA is advancing industrial autonomy by turning simulated experience into real-world capability. This hands-on ML engineering role focuses on solving sim-to-real challenges, owning the learning pipeline from simulation data generation to policy training, evaluation, and deployment on physical robots.

You will collaborate with AI, controls, hardware, and robot operations teams to improve fidelity, variation, and verification, applying imitation learning and RL to

Qualifications

• Demonstrated ability to improve sim-to-real transfer by changing the simulation, data-generation process, or training distribution.
• Hands-on experience training or adapting VLA policies, WAMs, or other visuomotor policies, with strong practical understanding of imitation learning and RL.
• Experience building simulation environments or pipelines for robot learning using tools like Isaac Sim / Isaac Lab, MuJoCo, robosuite, or comparable systems.

Responsibilities

• Lead simulation-side sim-to-real development and identify gaps in observations, physics, contact, and control.
• Build and maintain simulation environments and data-generation pipelines for robot learning.
• Train and adapt visuomotor policies using simulated and real robot data; design reproducible experiments.

Skills

Python

PyTorch

Robotics

C++

Reinforcement Learning

Education

MS in robotics

PhD in related field

Tools

Isaac Sim

MuJoCo

robosuite

Job description

Noble Machines in Sunnyvale, CA is advancing industrial autonomy by turning simulated experience into real-world capability. This hands-on ML engineering role focuses on solving sim-to-real challenges, owning the learning pipeline from simulation data generation to policy training, evaluation, and deployment on physical robots.

You will collaborate with AI, controls, hardware, and robot operations teams to improve fidelity, variation, and verification, applying imitation learning and RL to