You will join an AI infrastructure team that builds tools to make agents more reliable in production.
The role is backend-leaning but spans full-stack work across product interfaces and platform systems.
Your work turns real agent failures found in the field into workflows and product changes.
Responsibilities
Build product interfaces for agent investigations, evaluation, trajectory debugging, verification, and replay.
Take ownership of problems from start to finish: talk with customers about where their agents fail, decide on the fix, and keep improving features after release.
Develop APIs, data models, internal systems, and user-facing surfaces across the whole stack.
Build systems that can run large-scale investigations over production traces.
Design platform features such as workspaces, permissions, billing, usage tracking, SDKs, and terminal-first developer workflows.
Work closely with AI teams to learn what goes wrong with agents in production and shape product decisions based on it.
Requirements
Typically 5 to 8 years of full-stack product engineering experience, weighted toward backend work.
You have shipped polished, user-facing products end to end, including frontend work in React or Next.js and backend development.
You are strong in TypeScript and React, plus Python or Node.js, and have designed APIs, modeled data, and debugged production systems.
You have worked directly with customers and turned what they need into shipped product.
Experience with AI agents, LLM-powered products, agent architectures, evaluation or debugging tools, or developer tooling is beneficial but not required.
Conditions
Base salary is $230,000 to $300,000 per year.
Visa sponsorship is available.
The role is on-site in San Francisco, California, United States.