AI Engineer, Simulation Environments
On-site · Munich (Germany) · Mid · Full-time
Want offers like this one?
Create your free PyHires profile: you will see how well you match each offer and get a weekly email with the ones that fit you.
Overview
- Work mode
- On-site
- Location
- Munich (Germany)
- Engagement
- Full-time
- Seniority
- Mid
- Languages
- English
- Open to candidates from
- EU countries only
- Company
- Vesoma
- Posted
- Nov 19, 2026
About the role
Vesoma is building humanoid robots for physical labour, engineered to automotive standards, with its two main locations in Munich and Limassol. You will design and build the simulation environments, assets and benchmarks used to train and evaluate its humanoid robot policies, and keep simulation aligned with reality. This is a full-time position in the AI team.
What you'll do
- Build scalable simulation environments for humanoid robots with physics engines such as MjLab and Isaac Lab, with training tasks, curricula and evaluation scenarios
- Build and curate the database of robot, object and scene assets: import, validation, collision and physics properties, metadata and versioning
- Own large-scale simulation benchmarks and evaluation sweeps with sound metrics and regression tracking
- Align simulation with reality through robot modelling, actuator and sensor models and system identification
- Turn real customer use cases into simulated environments and benchmarks
What you bring
- 3+ years building simulation environments for robot learning with MuJoCo/MjLab, MuJoCo Warp/MJX, Isaac Lab/Isaac Sim or Newton
- Rigid-body dynamics and contact simulation, including solver parameters and numerical stability
- Robot and scene formats and asset pipelines: URDF, MJCF, OpenUSD, mesh processing and convex decomposition
- GPU-parallel simulation, sweep orchestration and reproducibility
- Core stack: Python, C++, PyTorch and/or JAX, and CUDA or NVIDIA Warp
Nice to have
- Video game development (Unreal Engine, Unity or Godot)
- Sim-to-real techniques such as system identification, actuator modelling and domain randomisation
- Photorealistic rendering and scene reconstruction, Blender or Houdini
- RL and imitation learning for humanoid locomotion and manipulation
Requirements
Based on:Python
Must-have
- C++
- JAX
- PyTorch
About Vesoma
Vesoma is building humanoid robots for physical labour, engineered to automotive standards of safety, scale and durability, with the goal of becoming a European leader in physical AI. It is well capitalised and has two main locations, in Munich and Limassol.
Stack
- Python
- C++
- JAX
- PyTorch
Company
Vesoma
Germany