SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mecka AI is building the data infrastructure layer for robotics and embodied AI. The company designs and operates global systems for data capture, data labeling, and hardware-enabled workflows used by leading AI labs and robotics companies to train and validate humanoid and embodied AI systems.
As a Research Engineer in Simulation & World Models, you will build simulation systems that train and evaluate robots and embodied agents. This is a hands-on role focused on modeling physical interactions, creating scalable training worlds, and connecting simulation with real-world data across Mecka Labs. You will work closely with frontier robotics teams to bridge real-world data, simulation, learning-based systems, and deployed hardware.
Key responsibilities include:
- Building simulation environments that model robots, sensors, objects, contacts, materials and dynamics, grounded in real behavior
- Scaling training workloads with reliable pipelines for parallel rollouts, synthetic data, policy training and evaluation
- Improving physical fidelity by calibrating against measured data and making targeted improvements to dynamics and sensing
- Driving sim-to-real transfer using system identification, domain randomization and controlled experiments
- Developing world models by building learned dynamics or latent models and combining them with physics-based systems
- Collaborating with researchers and engineers to turn prototypes into reusable simulation assets and documented methods
You will own whether the simulated world behaves credibly and runs at scale; partner roles will turn it into trainable tasks and reliable evaluations.
REQUIREMENTS:
- Deep experience building and debugging physics-based simulation with MuJoCo, Isaac Sim or comparable platforms
- Strong reinforcement learning and embodied AI fundamentals, including observations, actions, rewards and policy evaluation
- Experience with learned dynamics, latent world models or hybrid physics-learning systems
- Strong Python and C++ systems skills; ability to build scalable, reproducible tools that other researchers can extend
- Ability to design controlled experiments, measure sim-to-real gaps and trace failures to models, data, policies or infrastructure
PREFERRED QUALIFICATIONS:
- Simulation or training infrastructure used at scale by a robotics or embodied AI research team
- Demonstrated sim-to-real transfer in manipulation, locomotion, navigation or another physical domain
- Published or open-source work in world models, differentiable simulation, GPU-accelerated simulation or model-based reinforcement learning
