Staff Infrastructure Engineer: Frontier AI & Robotics

Amazon Fulfillment Technologies & Robotics; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Amazon’s Frontier AI & Robotics (FAR) team is seeking a Member of Technical Staff, Infrastructure to design and operate the foundational systems powering robotics research and development. You will build scalable infrastructure that enables model training, large-scale experiments, and robotic deployments at Amazon scale in San Francisco.

The role emphasizes performance, reliability, and low latency, designing systems to handle vast training data and demanding compute tasks while delivering a

Responsibilities

• Design and build scalable compute and data infrastructure to support model training, inferencing, and eval for frontier AI/Robotics development
• Lead large technical initiatives and shape the architecture of FAR’s research platform infrastructure
• Develop tooling and frameworks that accelerate research workflows, including dataset management, visualization, and quality assessment systems
• Optimize query performance and data availability for experimentation and analytics workflows used by research teams
• Improve the performance, efficiency, and reliability of FAR’s core compute and storage infrastructure, ensuring systems remain fast and stable at scales
• Build highly scalable experimentation and analytics infrastructure to support model evaluation, A/B testing, and feature performance
• Collaborate directly with science and robotics teams to support research projects through both infrastructure development and hands‑on technical contribution

Job description

Amazon’s Frontier AI & Robotics (FAR) team is seeking a Member of Technical Staff, Infrastructure to design and operate the foundational systems powering robotics research and development. You will build scalable infrastructure that enables model training, large-scale experiments, and robotic deployments at Amazon scale in San Francisco.

The role emphasizes performance, reliability, and low latency, designing systems to handle vast training data and demanding compute tasks while delivering a