AI Agent Engineer (Coding Agent)

Role overview Own the agent infrastructure that generates interactive apps, improving how autonomously, quickly, and reliably the system can build software. The role spans agent workflow design, evaluation, model selection, and production operations, with responsibility for taking new approaches from research into use.

Responsibilities - Design and maintain the harness and orchestration layer for long-running agent workflows, including tool use, context handling, and retry or repair paths. - Build evaluation systems, regression checks, and failure classifications to track and improve generation quality. - Select and benchmark models, balancing reliability, cost, and latency. - Make agent behavior diagnosable through tracing, metrics, alerts, and other observability practices. - Research promising techniques and ship them into production.

Requirements - Experience building agentic systems using provider APIs or frameworks, with judgment about when frameworks help or hinder. - Strong software engineering skills, especially in Python and/or TypeScript. - Experience with distributed or real-time systems and operating production software. - Evidence of owning technically difficult work end to end and delivering measurable outcomes. - An experimental, research-minded approach paired with the ability to ship. - Product judgment to translate user experience goals into system design decisions.

Nice to have - Experience building a customer-facing product from its earliest stage. - Work on coding agents, developer tools, IDE assistants, or code-generation pipelines. - Experience with automated evaluation, regression suites, or release gates; LLM tool protocols; or model training and fine-tuning.

Benefits and work setup Remote-first work, competitive salary and equity, and health insurance coverage.