Staff ML Research Scientist - Robotics

Vancouver 3 days agoFull-time External
1m - 1.5m / yr
Staff Research Scientist – Robotics & Machine Learning Location: Vancouver (Hybrid – at least 2 days onsite) Compensation: $200–300k base + equity A pioneering robotics and AI team is hiring a Staff Research Scientist to push the boundaries of embodied intelligence. In this role, you’ll work end to end across research, simulation, and real-world deployment, developing ML-driven control systems for dexterous humanoid robots. You’ll shape the direction of reinforcement and imitation learning research, design scalable training and data pipelines, run large-scale simulation experiments, and bring cutting-edge algorithms onto robots equipped with advanced sensing capabilities. This is a rare opportunity to contribute research that transitions directly into physical, real-world performance. What You’ll Do • Develop and advance RL/IL algorithms for complex manipulation tasks • Lead high-impact research directions and guide experimentation strategy • Design training and data-collection systems that speed up deployment • Improve efficiency, reliability, scalability, and compute performance of learning systems • Run experiments in simulation and on real robots, debugging end to end • Diagnose failures and design robust, production-ready solutions • Collaborate closely with ML researchers, robotics engineers, and applied teams What We’re Looking For • PhD in ML, Robotics, Computer Science, Applied Math, or equivalent expertise • 5+ years building and deploying robotic manipulation systems (simulation + real robots) • 5+ years hands-on experience with RL/IL for real-world robotics • 4+ years developing large-scale parallel simulation pipelines • Strong background in continual learning and sim-to-real transfer • Experience transitioning ML research into production robotic systems • Publication record in top-tier robotics/AI venues (ICRA, IROS, CoRL) Technical Strengths • Expert Python • Proficiency with PyTorch or TensorFlow • Experience with ROS2 and parallel compute (CUDA, OpenMP) • Deep understanding of RL theory and practical implementation