Robotics Learning Engineer

Alexander Chapman Ltd; California; Presencial; USD 120 000 - 180 000; Tempo integral

Resumo da oferta

Alexander Chapman Ltd. in New York City is seeking a Simulation & Robot Learning Engineer to own the pipeline from high-fidelity simulation to policies deployed on real robots.

You will lead the design and tuning of full simulation environments, physics, scenes, and manipulation tasks. You will train robot policies using RL, imitation learning, diffusion policies, VLAs, and/or world models, and work to close the sim-to-real gap through domain randomization, system identification, calibration,

Qualificações

• 4+ years hands-on experience building simulation environments and deploying learned policies.
• Experience with Isaac Sim, MuJoCo, Drake, Gazebo, SAPIEN, Genesis or similar.
• Strong Python, PyTorch or JAX, and C++ for near-system work.

Responsabilidades

• Build and tune simulation environments, assets, physics, scenes, and manipulation tasks.
• Train robot policies using RL, imitation learning, diffusion policies, VLAs, and/or world models.
• Close the sim-to-real gap through domain randomization, system identification, calibration, and real-world debugging.
• Build scalable synthetic data, teleoperation, procedural generation, and automated curriculum pipelines.
• Use agentic AI to automate environment, task, reward, evaluation, and data generation.
• Develop benchmarking infrastructure that makes simulation performance predictive of hardware performance.

Conhecimentos

Python

PyTorch

JAX

C++

Reinforcement learning

Imitation learning

Domain randomization

World models

Ferramentas

Isaac Sim

MuJoCo

Drake

Gazebo

SAPIEN

Genesis

Descrição da oferta de emprego

In this role, you'll:

• Build and tune simulation environments, assets, physics, scenes, and manipulation tasks
• Train robot policies using RL, imitation learning, diffusion policies, VLAs, and/or world models
• Close the sim-to-real gap through domain randomization, system identification, calibration, and real-world debugging
• Build scalable synthetic data, teleoperation, procedural generation, and automated curriculum pipelines
• Use agentic AI to automate environment, task, reward, evaluation, and data generation
• Develop benchmarking infrastructure that makes simulation performance predictive of hardware performance

I'm partnering with an ambitious robotics company building and deploying humanoid robots in real factory environments. They're looking for a Simulation & Robot Learning Engineer to own the pipeline from high-fidelity simulation and large-scale data generation to policies that reliably transfer onto physical robots.

They're looking for someone with 4+ years of hands-on experience building simulation environments and deploying learned policies on real robots. Deep experience with Isaac Sim, MuJoCo, Drake, Gazebo, SAPIEN, Genesis, or similar is important, along with strong Python, PyTorch or JAX, and enough C++ to work close to the system. Experience with manipulation, deformable objects, bimanual systems, humanoids, GPU-accelerated simulation, or VLA/world models is a strong plus.

This is an on-site opportunity in New York City with significant ownership over the simulation and learning stack and the chance to solve some of the hardest problems in getting learned robotic behavior to work reliably in the real world.