Robotics Evaluation Scientist — Teleoperation & Data

Amazon Inc.; Seattle, Northern (WA, KY); Hybrid; USD 172,000 - 223,000; Full time

Job summary

Amazon.com Services LLC in New York is seeking an Applied Scientist to lead evaluation frameworks for robotic capabilities and data collection protocols across real-world tasks. You will design measurement, stress-testing, and improvement of robot behavior to accelerate system performance.

Work spans robotics, ML, and human-in-the-loop systems, building infrastructure for evaluation policies, task definitions, and operator-facing interfaces that enable scalable data collection and reliable

Qualifications

• 3+ years building models for business applications.
• Publications or patents in top-tier venues.
• Proficiency in Java, C++, Python.
• Experience with algorithms, data structures, and HPC.

Responsibilities

• Design and implement evaluation frameworks to measure robot capabilities across real-world scenarios.
• Develop task definitions, success criteria, and benchmarking methodologies for evaluation of policies.
• Create and refine data collection protocols to generate high-quality task-relevant datasets.
• Build and iterate teleoperation workflows and operator interfaces for scalable data collection.
• Analyze evaluation results to identify performance gaps and data collection opportunities.
• Collaborate with engineering to integrate evaluation tooling and data pipelines into the robotics stack.
• Stay current with robotics and human-in-the-loop learning to improve approaches.
• Lead technical projects from conception to production deployment.
• Mentor junior scientists and engineers.

Skills

Programming in Java/C++/Python

Publications or patents

Algorithms & HPC

Tools

Unix/Linux

Software development

Distributed computing

Job description

Amazon.com Services LLC in New York is seeking an Applied Scientist to lead evaluation frameworks for robotic capabilities and data collection protocols across real-world tasks. You will design measurement, stress-testing, and improvement of robot behavior to accelerate system performance.

Work spans robotics, ML, and human-in-the-loop systems, building infrastructure for evaluation policies, task definitions, and operator-facing interfaces that enable scalable data collection and reliable