Summary

Engineers will develop advanced motor controllers using reinforcement or imitation learning to enable natural, resilient, and adaptable robotic movements.
This involves close teamwork across disciplines to merge control systems with physical robot hardware, considering factors like actuator performance and energy use.

Tasks include refining and proving control methods using both simulations and actual hardware.
Individuals should remain current with progress in AI, control theory, and robotics to tackle complex motion problems.

Key duties involve leading technical projects from initial idea to full operational use, alongside guiding less experienced team members.
Furthermore, translating research breakthroughs into usable engineering products is a core responsibility.

The team focuses on creating safe, capable, and engaging robots for daily human interaction.
They design and train RL policies for various legged movements, such as walking, running, and recovering from falls on physical robots.

A major effort is placed on creating robust simulation-to-reality transfer methods to overcome discrepancies between digital models and the real world.
This includes combining control-based techniques with RL, using safety measures in training, and integrating RL outputs with full-body control.

The role requires building and maintaining extensive training platforms, including physics simulations and GPU parallelization for locomotion learning.
Investigations will explore distilling motion policies and integrating them with world models, foundation models, and perception systems.

Performance assessment is rigorous, utilizing simulation benchmarks, physical hardware testing, and detailed failure analysis.
The work involves publishing high-quality research and collaborating with perception and planning groups for environment-aware movement.

Title

Amazon Science homepage

Description

Learn about Amazon's scientific research, science community, and career opportunities in artificial intelligence (AI), machine learning (ML), computer vision, robotics, quantum, economics and more.

Keywords

research, amazon, university, learning, will, team, systems, machine, scientists, work, science, scale, customers, technical, teams, design, robotics

Categories

Technology

Upstreams

amazon.ca, amazon.com.tr, amazon.es

Downstreams

x.com, instagram.com, youtube.com, linkedin.com, github.com, amazon.jobs

NS Lookup

A 13.224.83.97, A 13.224.83.23, A 13.224.83.12, A 13.224.83.122

Dates

Created 2024-02-12

Updated 2026-08-10

Summarized 2026-09-04