Applied Scientist - ML and Robotics

About the role

At Amazon Robotics, we design advanced robotic systems capable of intelligent perception, learning, and action alongside humans, at massive scale. Our mission is to deploy robots that increase productivity and efficiency across Amazon fulfillment centers while operating safely and robustly in complex, contact-rich environments. We are seeking an Applied Scientist to develop manipulation controllers for robotic systems operating in contact-rich, uncertain environments.

Responsibilities

• Research, design, implement, and evaluate machine learning–based manipulation policies for contact-rich tasks, integrating learning with feedback control, estimation, and motion planning.
• Develop learning frameworks that leverage simulation, real-world data, and hybrid physics- and data-driven models to enable robust agency interaction, grasping, insertion, and object handling.
• Design and execute experiments in simulation and on hardware to train, validate, and stress-test learned manipulation policies under real-world variability and uncertainty.
• Collaborate with software engineering teams to deliver scalable, real-time, and maintainable implementations of learning-based manipulation algorithms in production robotic systems.
• Partner with cross-functional teams across perception, hardware, systems engineering, science, and operations to transition learned policies from research prototypes to reliable, production-ready capabilities across Amazon Robotics platforms.

Requirements

PhD, or Master's degree and 4+ years of science, technology, engineering or related field experience

Experience in patents or publications at top-tier peer-reviewed conferences or journals

Experience programming in Java, C++, Python or related language

Experience designing, running, and analyzing experiments in simulation and on real robotic hardware

Qualifications

• Strong foundation in robot dynamics, control, and state estimation, and experience integrating these with data-driven methods
• Hands-on experience with reinforcement learning, imitation learning, or hybrid learning–control approaches applied to robotics
• Familiarity with simulation tools and sim-to-real transfer for robotic manipulation
• Experience collaborating with software engineering teams to transition research prototypes into scalable, real-time production systems