Staff Engineer: Agentic AI and Evaluation Systems

Waymo; San Francisco (CA); Hybrid; USD 251,000 - 310,000; Full time

Job summary

Waymo is seeking an experienced data-minded software engineer or data scientist to advance evaluation in the Planner Evaluation team. This hybrid role reports to a Senior Engineering Manager and focuses on building robust AI evaluation frameworks and production-ready agent-based systems for autonomous driving software.

You will design, implement, and scale complex AI evaluation pipelines, collaborate with multiple teams, and help push the boundaries of safe, reliable autonomous decision making.

Qualifications

• 7+ years of software engineering experience in large-scale systems.
• Design, build, evaluate, and deploy agent-based AI solutions into production.
• Deep AI/ML expertise with multi-agent orchestration and LLM integration.
• Ability to define technical strategy and influence senior leadership.
• Strong communication and collaboration across diverse audiences.

Responsibilities

• Design, build, and deploy production-grade agentic AI systems to solve complex, open-ended evaluation challenges.
• Develop robust, automated evaluation frameworks and objective metrics to measure accuracy, safety, and reliability of non-deterministic AI agents.
• Automate building and scaling of the hardest eval problem inhibiting Waymo’s scale.

Skills

Agent-based AI

Production-grade services

Large-scale systems

Technical leadership

Communication

Education

Bachelor's degree in Computer Science or related field

PhD in CS or AI (preferred)

Tools

C++

Job description

Waymo is seeking an experienced data-minded software engineer or data scientist to advance evaluation in the Planner Evaluation team. This hybrid role reports to a Senior Engineering Manager and focuses on building robust AI evaluation frameworks and production-ready agent-based systems for autonomous driving software.

You will design, implement, and scale complex AI evaluation pipelines, collaborate with multiple teams, and help push the boundaries of safe, reliable autonomous decision making.