Robot Learning Engineer

Maxwell Bond; New York (NY); On-site; USD 120,000 - 180,000; Full time

Job summary

Maxwell Bond is seeking a Member of Technical Staff (Simulation & Robot Learning Engineer) to build high-fidelity simulation environments and develop scalable data-generation pipelines for robot learning.

You will train policies using reinforcement learning, imitation and diffusion methods, and work on sim-to-real transfer to real humanoid robots. This hands-on role covers simulation, ML, robotics and hardware on challenging manipulation tasks.

Qualifications

• 4+ years building simulation environments and training robot-learning policies.
• Experience transferring policies from simulation to real hardware.
• Strong Python and C++ programming skills.
• Familiarity with RL, imitation and diffusion-based methods.

Responsibilities

• Build high-fidelity simulation environments.
• Create scenes, assets, and physics configurations.
• Train deployable robot policies.
• Transfer policies to physical robots.
• Close sim-to-real gap via domain randomization and calibration.
• Build scalable data pipelines for synthetic data and demonstrations.
• Develop evaluation and benchmarking infrastructure.
• Apply agentic AI and LLM-based tooling to automate workflows.
• Collaborate with controls, perception, mechanical and learning engineers.

Skills

Robot learning

Python

C++

PyTorch

JAX

Reinforcement learning

Imitation learning

Diffusion policies

Vision-Language-Action

World models

Education

BS/MS/PhD in Robotics, CS, ML, EE/Mech Eng

Tools

Isaac Sim/Lab

MuJoCo

Drake

Gazebo

SAPIEN

Genesis

Job description

Our client is building and deploying humanoid robots designed to operate on real factory floors and perform meaningful work in demanding production environments.

They are looking for a Member of Technical Staff (Simulation & Robot Learning Engineer) to build high-fidelity simulation environments, develop scalable data-generation pipelines, train robot-learning policies, and close the sim-to-real gap so that behaviors developed in simulation reliably transfer to physical robots.

You will own the pipeline from simulation and task authoring through data generation, policy training, evaluation and deployment, working across reinforcement learning, imitation and diffusion policies, Vision-Language-Action models, world models and emerging AI techniques.

This is a hands-on individual contributor role spanning simulation, machine learning, robotics and physical hardware, with a focus on challenging manipulation problems across single-arm, bimanual, whole-body and humanoid platforms.

What You'll Do

• Build and maintain high-fidelity simulation environments using platforms such as Isaac Sim/Lab, MuJoCo, Drake, Gazebo, SAPIEN, Genesis or similar.
• Create scenes, assets, physics configurations, contact models and robotic manipulation tasks.
• Train deployable robot policies using reinforcement learning, imitation learning, diffusion policies, VLAs and related approaches.
• Transfer policies from simulation to physical robots and improve real-world robustness.
• Close the sim-to-real gap through domain randomization, system identification, calibration and physics tuning.
• Build scalable pipelines for synthetic data, teleoperation, demonstrations, procedural scene generation and automated training.
• Develop evaluation and benchmarking infrastructure for simulation and real-world performance.
• Apply agentic AI and LLM-based tooling to automate simulation, data and training workflows.
• Work closely with controls, perception, mechanical and learning engineers to deploy reliable robot behaviors.

What We're Looking For

• BS, MS or PhD in Robotics, Computer Science, Machine Learning, Electrical/Mechanical Engineering or a related field.
• 4+ years of hands-on experience building simulation environments and training or deploying robot-learning policies on physical hardware.
• Strong experience with at least one major robotics simulator, such as Isaac Sim/Lab, MuJoCo, Drake, Gazebo, SAPIEN or Genesis.
• Experience building and tuning simulation environments rather than only using existing ones.
• Demonstrated experience with sim-to-real transfer, including domain randomization, system identification and calibration.
• Strong understanding of modern robot learning, particularly reinforcement learning, imitation or diffusion-based policies.
• Familiarity with Vision-Language-Action models and/or world models.
• Experience with robot-learning data generation, teleoperation or demonstration pipelines.
• Strong Python skills and familiarity with C++.
• Experience with PyTorch, JAX or comparable ML frameworks.
• Experience getting robot-learning systems working reliably on physical hardware.
• Experience with manipulation using robotic arms, bimanual systems, humanoids or similar platforms.

Nice to Have

• Experience with agentic AI or LLM-based simulation and data workflows.
• Deformable-object manipulation, including cloth, cables or soft materials.
• Experience with VLA systems or robotics world models.
• Bimanual, whole-body or wheeled-humanoid manipulation experience.
• GPU-accelerated or large-scale parallel simulation.
• Large-scale reinforcement-learning training.
• Familiarity with robotics controls and hardware stacks.
• Publications, open-source contributions or other demonstrated work in robot learning or simulation.

What We're Looking For

The ideal candidate combines strong robotics and machine-learning fundamentals with practical engineering experience. You should enjoy building systems from the ground up, scaling experiments and solving the difficult problems that arise when moving from simulation to physical robots.

This is an opportunity to work across high-fidelity simulation, large-scale data generation, modern robot learning and real-world humanoid robotics.

Our client welcomes applicants from all backgrounds and is committed to building an inclusive and equal-opportunity workplace.