Das ist der Job
Push large-scale physics simulation to the point where self-play replaces teleoperation.
Darum lohnt es sich
Co-design simulation interfaces with the RL and world-model teams. To get there, we are assembling a small, highly selective founding team of the strongest researchers and engineers we can find. Compensation & benefits Base salary €100,000+ , calibrated to seniority and impact.
Founding-team seat in a stealth lab, based at Munich Urban Colab. #J-18808-Ljbffr Mission. Self-play scales only as far as our simulator does. You make contact-rich physics fast, parallel and predictive enough for foundation-model-scale RL. What you'll do Extend GPU physics simulators (Isaac, MuJoCo MJX, Genesis, custom) for cross-embodiment.
Close the sim-to-real gap with domain randomisation and system identification. What we look for Strong background in rigid-body / contact-rich simulation. C++ and Python; comfortable with CUDA or GPU compute. Shipped large-scale robotic simulation environments. Nice to have Differentiable simulation or neural rendering experience.
Published work on sim-to-real transfer or domain randomisation. Why omn4i omn4i is on a mission to become Europe's leading AI lab for general-purpose robotics . We are building one physics-grounded world model for every robot — humanoid, quadruped, wheeled, aerial — and we believe the GPT-moment for robotics will happen this decade.
We want it to happen in Europe. No politics, no legacy stack, no committee science — just frontier work, shipped fast, with people who have done it before. Every early hire is an owner, not an employee. Meaningful company shares — every early hire is an owner. Top-tier compute budget and the hardware you need.
