Applied Scientist - ML and Robotics
Core
Design force-aware control strategies and data-driven manipulation policies for robots operating in contact-rich, uncertain environments to enable robust grasping, insertion, and object handling.
Role type
Applied Scientist (Robotics & Control)
Builds
Manipulation controllers and learning-based policies for Amazon Robotics platforms
Domain
Robotics, Control Theory, Machine Learning
Deliverable
production ML models
Required skills
robot dynamics, control theory, state estimation, reinforcement learning, imitation learning, simulation-to-real transfer, C++/Python/Java programming, experimental design in simulation and on hardware
Preferred skills
manipulation policy development for contact-rich tasks, hybrid physics-data-driven modeling, real-time system implementation, cross-functional collaboration
Technologies
C++, Python, Java, simulation tools
Responsibilities
Research and implement ML-based manipulation policies integrating learning with feedback control; develop learning frameworks leveraging simulation and real-world data; design and execute experiments to validate policies on hardware; collaborate with software engineering to deploy scalable algorithms; partner with cross-functional teams to transition prototypes to production.
Seniority
Senior, hands-on IC