Applied Scientist - Simulation and Large-Scale RL, Amazon Robotics - Vulcan Stow

About the role

Our organization in Amazon Robotics builds robots that perform contact-rich manipulation safely and reliably in complex, unstructured environments at Amazon scale. We are seeking an Applied Scientist to advance reinforcement learning for manipulation. The goal is to create robots that learn how to push, flip, rearrange, and dexterously insert items with unparalleled robustness, speed, and reliability, deployable across Amazon's global network to handle the full diversity of items Amazon sells.

You will focus on training these policies in simulation at scale and transferring them to physical robots. You will join a small team whose mission reaches beyond any single product: to invent and apply manipulation capabilities that generalize to many future robotics applications. The robots our organization already deploys at scale provide a rare proving ground to collect data, run experiments, and get new policies onto real hardware faster than almost anywhere in the field.

Responsibilities

• Design and train reinforcement learning policies for non-prehensile and contact-rich manipulation, focused on the long tail of diverse, demanding conditions.
• Build and scale simulation environments, training curricula, and reward formulations that produce policies which transfer to real robots.
• Develop sim-to-real methods (domain randomization, system identification, and related techniques) and validate them on physical hardware.
• Write production-quality code and own scalable, efficient training and evaluation pipelines.
• Evaluate policy behavior and failure modes, and iterate between simulation and real-world testing to improve robustness.
• Partner with scientists and engineers across control, perception, and hardware to move ideas from prototype to demonstrated capability.
• Represent Amazon Robotics in academia through publications and scientific presentations.

Requirements

• PhD, or Master's degree and 4+ years of science, technology, engineering or related field experience.
• Experience programming in Java, C++, Python or related language.
• Experience in patents or publications at top-tier peer-reviewed conferences or journals.
• Experience training reinforcement learning or imitation learning policies for manipulation or robot control problems.
• Strong background in reinforcement learning, including reward design and training at scale in simulation.

Preferred Qualifications

• Experience in professional software development.
• Experience with sim-to-real transfer (domain randomization, system identification, etc.), including transferring learned policies onto physical robots.
• Familiarity with learned dynamics or world models for manipulation.

Benefits

The benefits that generally apply to regular, full-time employees include:

• Medical, Dental, and Vision Coverage.
• Maternity and Parental Leave Options.
• Paid Time Off (PTO).
• 401(k) Plan.

Pay

USA, WA, Seattle - $142,800.00 - $193,200.00 USD annually. Your Amazon package will include sign-on payments and restricted stock units (RSUs). Final compensation will be determined based on factors including experience, qualifications, and location.