Simulation Engineer - Manipulation

Here at Humanoid, we believe in a future where robots amplify human potential. That’s why we’ve set out on a mission to build the world’s most capable, commercially-scalable, and safe humanoid robots. We’re bringing that mission to life with HMND‑01 - our rapidly developed humanoid platform being deployed in real industrial environments - and we’re growing the team to take it even further.

We're hiring a Senior Simulation Engineer to join our VLA team based in London.

In this role, you will build and maintain simulation environments for dexterous manipulation tasks across industrial, service, and home domains. This is primarily a simulation and reinforcement learning-focused role, so we are looking for experience creating realistic physics-based environments and training RL policies, while experience in robotics isn't strictly required. However, if you don't have such experience, be prepared that you'd need to familiarize yourself with a new domain quickly.

• →

  Design and implement gym environments for manipulation tasks spanning industrial, service, and home settings, defining appropriate observation spaces, action spaces, and reward functions.

• →

  Analyze and address reward hacking - identify cases where learned policies exploit reward misspecification and iterate on reward design to produce robust behaviors.

• →

  Analyze and reduce the sim-to-real gap by tuning physics parameters, improving asset fidelity, and validating simulation behavior against real-world data.

• →

  Work with third-party vendors to procure, validate, and integrate high-quality 3D assets (objects, fixtures, environments) suitable for physics-based simulation.

• →

  Ensure correct physics setup - contact dynamics, friction, mass properties, joint limits - so that trained policies transfer reliably to hardware.

• →

  Identify and reduce simulation bottlenecks to maximize training throughput and environment step rates.

• →

  Improve our simulation-based evaluation and reinforcement learning infrastructure to support rapid iteration and scaling.

• 3+ years building simulation environments or game-engine-based interactive systems (industry or research) with shipped products, published results, or equivalent artifacts to show for it.

• Deep hands-on experience with at least one physics simulator (Isaac Sim/IsaacLab, MuJoCo, PyBullet, Drake) or equivalent game engine experience (Unreal, Unity) with a focus on physically accurate interactions.

• Strong practical experience running large-scale parallel simulation on GPU clusters and good familiarity with modern GPU-accelerated simulation infrastructure.

• Strong Python; you can profile bottlenecks, debug physics issues, and write maintainable research code.

• Familiarity with modern software engineering practices.

• You document experiments clearly and communicate trade-offs crisply.

Nice to have:

• Robotics or manipulation-specific experience (grasping, contact-rich tasks, deformable objects).

• Experience designing reward functions and training RL policies in simulated environments; solid understanding of common failure modes (reward hacking, distribution shift, sim-to-real gap).

• Experience with NVIDIA Isaac Sim & IsaacLab specifically.

• Experience with domain randomization, system identification, or other sim-to-real transfer techniques.

• Publications at top-tier robotics or RL conferences or equivalent open-source contributions.

• Familiarity with robocasa, robosuite or similar open-source manipulation simulation frameworks.