Robotics Simulation & World Models Research Engineer

Mecka; New York (NY); On-site; USD 170,000 - 200,000; Full time

Job summary

Mecka AI seeks a research engineer to build simulation systems that train and evaluate robots and embodied agents. You will model physical interactions, create scalable training worlds, and connect simulation with real-world data across Mecka Labs.

This hands-on role covers using MuJoCo and Isaac Sim, improving fidelity, and pushing sim-to-real transfer through domain randomization and learned world models, with Python and C++ tooling for robust, reusable research assets.

Qualifications

• Deep experience building and debugging physics-based simulation.
• Strong reinforcement learning and embodied AI fundamentals.
• Experience with learned dynamics or hybrid physics-learning systems.
• Strong Python and C++ systems skills for scalable tools.

Responsibilities

• Build simulation environments modeling robots, sensors, objects, and dynamics.
• Scale training workloads and pipelines for parallel rollouts and data.
• Improve physical fidelity by calibration against real data.
• Drive sim-to-real transfer with domain randomization and system identification.
• Develop world models and test their value for prediction and control.
• Collaborate with researchers and engineers to turn prototypes into reusable assets.

Skills

Simulation physics

Reinforcement learning

World models

Python

C++

Tools

MuJoCo

Isaac Sim

Python

C++

Job description

Mecka AI seeks a research engineer to build simulation systems that train and evaluate robots and embodied agents. You will model physical interactions, create scalable training worlds, and connect simulation with real-world data across Mecka Labs.

This hands-on role covers using MuJoCo and Isaac Sim, improving fidelity, and pushing sim-to-real transfer through domain randomization and learned world models, with Python and C++ tooling for robust, reusable research assets.