Software Engineer - Simulation
AI Summary
Simulation Engineer focused on building and scaling physics simulation infrastructure for robotics, including sim-to-real transfer, domain randomization, and training environments.
About this role
Join Us in Building the Future of Physical AI
We build the simulation infrastructure for physical AI. We develop GenAI-powered tools that enable robotics teams to create unlimited, diverse training data and realistic evaluation environments. Our team spun out of MIT CSAIL, where we pioneered techniques that trained robots using only synthetic data. We're helping robotics companies transition to AI-native development workflows.
We're a lean team of researchers and engineers from DeepMind, OpenAI, FAIR, and top universities including MIT, Berkeley, Caltech, Harvard, and Yale. We've published best papers at top robotics conferences, won International Olympiad medals, and built core systems at leading AI labs. We believe in building complex systems that bring simplicity to our customers.
Global Team: We operate across US and China time zones. We value people who communicate proactively, document thoroughly, and take ownership of smooth handoffs across teams.
What to Expect
We're looking for a Simulation Engineer to build and scale our physics simulation infrastructure. You'll work on sim-to-real transfer, domain randomization, and creating training environments that enable robots to learn in simulation and perform in the real world.
What You'll Do
Build and maintain simulation environments using MuJoCo, PyBullet, and Isaac Lab
Develop sim-to-real transfer pipelines and domain randomization systems
Create scalable infrastructure for parallel simulation execution
Design RL training environments with realistic physics and diverse scenarios
Collaborate with ML researchers on environment design for policy learning
What You'll Bring
3-5 years of experience with physics simulation or robotics software
Proficiency with MuJoCo, PyBullet, Isaac Lab/Sim, or similar engines
Strong Python skills; familiarity with C++ for performance-critical code
Experience with reinforcement learning training pipelines
Understanding of robot dynamics, kinematics, and control
Nice to Have
Experience with sim-to-real transfer in deployed robotics systems
Background in domain randomization and synthetic data generation
Familiarity with GPU-accelerated simulation (Isaac Gym, Brax)
Publications or projects in robot learning
We believe diverse teams build better products. Even if you don't meet every requirement listed, we encourage you to apply if you're excited about this role and our mission.
