Principal Applied Scientist, Contact-Rich Manipulation, AR Manipulation HW & Systems
Core
Define control architectures for single- and dual-arm contact-rich manipulation to pick, grasp, and move millions of items in Amazon's fulfillment network.
Role type
Principal Applied Scientist (Robotics Control & Manipulation)
Builds
Robotic workcells for high-throughput logistics and fulfillment centers
Domain
Robotics, Control Theory, Machine Learning, Logistics
Deliverable
production ML models | product features
Required skills
Force/impedance/admittance control, Python, C++, integrating learned policies with low-level controllers, technical direction setting, sim-to-real transfer, tactile sensing integration, bimanual coordination
Preferred skills
Reinforcement/imitation learning, whole-body coordination, live industrial automation deployment, technical leadership via patents/publications
Technologies
Python, C++, Reinforcement Learning, Imitation Learning, Tactile Sensing, Simulation
Responsibilities
Own technical direction for contact-control strategy, identify open scientific problems, architect compliant-contact behaviors, partner with hardware/systems engineers, create mechanisms to learn from fielded production, establish reference implementations and evaluation standards, mentor scientists and engineers
Seniority
Principal, strategy & mentorship