DESCRIPTION

Amazon Lab126 is an inventive research and development company that designs and engineers high-profile consumer electronics. Lab126 began in 2004 as a subsidiary of Amazon.com, Inc., originally creating the best-selling Kindle family of products. Since then, we have produced innovative devices like Fire tablets, Fire TV, Amazon Echo, and the Astro household robot. What will you help us create?

As a Senior Reinforcement Learning Controls Scientist, you will be responsible for developing Reinforcement Learning models to control complex electromechanical systems. You will take responsibility for defining frameworks, performing analysis, and training models that guide and inform mechanical and electrical designs, software implementation, and other software modules that affect overall device safety and performance. You understand trade-offs between model-based and model-free approaches. You will demonstrate cross-functional collaboration and influence to accomplish your goals. You will play a role in defining processes and methods to improve the productivity of the entire team. You will interface with Amazon teams outside your immediate organization to collaborate and share knowledge. You will investigate applicable academic and industry research, prototype and test solutions to support product features, and design and validate production designs that deliver an exceptional user experience.

Key job responsibilities
- Produce models and simulations of complex, high degree-of-freedom dynamic electromechanical systems
- Train Reinforcement Learning control policies that achieve performance targets within hardware and software constraints
- Hands-on prototyping and testing of physical systems in the lab
- Influence hardware and software design decisions owned by other teams to optimize system-level performance
- Lead cross-functional teams (controls, firmware, perception, planning, sensors, mechanical, electrical, etc.) to solve complex system integration issues
- Define key performance indicators and allocate error budgets across hardware and software modules
- Perform root cause analysis of system-level failures and distinguish between hardware/software failures and hardware/software mitigations
- Translate business requirements to engineering requirements and identify trade-offs and sensitivities
- Mentor junior engineers in good design practice; actively participate in hiring of new team members

About the team
The Dynamic Systems and Control team develops models, algorithms, and code to bridge hardware and software development teams and bring robotic products to life. We contributed to Amazon Astro (https://www.amazon.com/Introducing-Amazon-Astro/dp/B078NSDFSB) and Echo Show 10 (https://www.amazon.com/echo-show-10/dp/B07VHZ41L8/), along with several new technology introductions and unannounced products currently in development.