Grafton Sciences is building physical general intelligence, and this role is to build high-fidelity simulation environments that power robotic policy training, safety validation, digital twins, and long-horizon planning.
Requirements
- Strong experience building robotic simulation environments, including physics modeling, contact dynamics, or manipulation simulation.
- Familiarity with domain randomization, calibration techniques, and methods for improving sim-to-real alignment.
- Ability to integrate simulation with ML pipelines, RL environments, and real hardware telemetry.
- Comfortable working across simulation tools, robotics engineering, and ML systems in fast-moving, interdisciplinary settings.
Responsibilities
- Build and maintain high-fidelity robotic simulation environments using MuJoCo, Isaac Gym, Drake, or equivalent frameworks.
- Develop domain randomization, parameter perturbation, and sensor/actuator modeling strategies to improve sim-to-real transfer.
- Integrate simulation with digital twins, RL environments, and evaluation pipelines for policy development and safety testing.
- Collaborate closely with robotics, controls, RL, and digital twin teams to align simulation fidelity with real-world system behavior.
Other
- Above all, we look for candidates who can demonstrate world-class excellence.
