Sim2Real Robotics Research Scientist
Core
Develop, deploy, and debug reinforcement learning (RL) policies for dexterous robotic manipulation on novel hardware, bridging the gap between high-fidelity simulation and real-world execution.
Role type
Research Scientist (Sim2Real Robotics)
Builds
Dexterous manipulation systems for complex manufacturing and construction tasks
Domain
Robotics, Deep Reinforcement Learning, Sim2Real Transfer
Deliverable
production ML models
Required skills
Deep reinforcement learning, Sim2Real transfer methodologies, multi-modal sensor integration (vision, tactile, force/torque), curriculum training design, physics simulation engines, Python, hardware deployment
Preferred skills
NVIDIA IsaacLab framework, locomotion/whole-body loco-manipulation expertise, domain randomization, real2sim alignment
Technologies
Python, Physx, MuJoCo, Netwon, NVIDIA IsaacLab
Responsibilities
Develop and debug RL policies on real-world robotic hardware; Investigate and mitigate domain gaps between simulation and reality; Design curriculum training and reward engineering strategies; Integrate multi-modal sensor inputs into real-time architectures; Explore policy learning for dynamic contact-rich action spaces; Develop high-throughput experimental workflows.
Seniority
Senior, hands-on IC
