Software Engineer, DevAI
Overview:
In this role you will help build the data and evaluation flywheel for large-scale, agentic software engineering. You will drive intelligence, reliability, and autonomy of coding agents within a multi-repo, AI-driven environment. You collaborate with distributed teams to shape technical roadmaps and oversee project execution, while nurturing a culture of sustainable innovation. Your work directly impacts automated refactoring, optimization, and agent performance at scale, balancing customer needs with product-area goals. This is a chance to contribute to Google's next generation of AI-assisted coding tools and workflows. Compensation / Benefitsbonus targetequitybenefits
Responsibilities:
- Build and improve data and evaluation flywheels for agentic coding systems
- Transform execution trajectories into actionable loss analysis to guide improvements
- Develop automated skill-improvement workflows and hillclimbing for agent success rates
- Evaluate and maintain coding agents across multi-repo tasks (refactoring, proto optimization, dependency pruning, Mendel flag deprecation)
- Focus on AI verifier, quality hillclimbing, and trajectory analysis within coding agents
- Collaborate with distributed teams to define roadmaps and ensure alignment and delivery
Key requirements:
- Bachelor in any field or equivalent practical experience 2 years of Python coding experience 2 years of experience with Generative AI2 years of experience with Large Language Models
- Experience with Agentic AI, evaluation, and quality hill climbing
- Experience with prompt engineering; using, interacting, and evaluating AI Agents or LLMs in coding space
- Cross-functional collaboration
- Ambiguity tolerance and rapid iteration
- Psychological safety and inclusive teamwork
- PythonGenerative AILarge Language Models