Robotics Simulation Research Engineer

Mecka; New York, Northern (NY, KY); Hybride; USD 140 000 - 210 000; Plein temps

Résumé du poste

Mecka in New York, NY, is seeking a hands-on 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.

You\'ll work with MuJoCo, Isaac Sim and related tools, bridging sim-to-real gaps, developing learned or hybrid world models for prediction, planning and control, and turning prototypes into reusable simulation assets with

Qualifications

• Proficient in building physics-based simulations and debugging complex systems.
• Strong foundation in robotics learning, RL, and embodied AI concepts.
• Experience with world models or hybrid physics-learning approaches.
• Solid Python and C++ software engineering skills for scalable tools.
• Ability to design controlled experiments and diagnose sim-to-real gaps.

Responsabilités

• Build simulation environments for robots, sensors, objects, contacts, materials, and dynamics.
• Scale training workloads with parallel rollouts, synthetic data, and policy evaluation.
• Improve physical fidelity through calibration against real data and targeted fixes.
• Drive sim-to-real progress using system identification and domain randomization.
• Develop learned dynamics or latent world models and integrate with physics-based systems.
• Collaborate with researchers to turn prototypes into reusable simulation assets.

Connaissances

Physics-based sim

Robotics learning

World models

Research engineering

Experimental judgment

Outils

Python

C++

Description du poste

Mecka in New York, NY, is seeking a hands-on 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.

You\'ll work with MuJoCo, Isaac Sim and related tools, bridging sim-to-real gaps, developing learned or hybrid world models for prediction, planning and control, and turning prototypes into reusable simulation assets with