Staff AI Engineer

Responsibilities

• You don't need to do all of these on day one, but these are the kinds of problems you'll own
• Set the technical direction for the agentic pipeline: Make our existing ticket-to-PR pipeline dependable infrastructure.
• Own its architecture, including agent orchestration, context management and tool calling, and improve its reliability, throughput, cost and developer experience across mobile, backend and web repositories. Run it unattended where the risk allows.
• Close the loop across the SDLC: Extend agentic workflows beyond code generation into spec quality checks, test generation, automated review, CI failure triage and repair, release verification, and post-deploy signals, so the system improves from what happens after merge.
• Prototype fast, then harden: Get a new workflow running in days, prove it on real tickets with real squads, and turn what holds up into production infrastructure.
• Run AI in production properly: Agents in production take more than an API call.
• Own the eval sets, LLM-as-a-judge grading, versioning, tracing and dashboards that tell us whether a workflow change made things better, including the hardest case: a change that passes CI and is still wrong. Track the numbers that matter (merge rate, rework, time-to-signal, cost per change, escaped defects) and make…
• Make autonomy safe
• Design sandboxing, permissions, human checkpoints and guardrails proportional to risk, with privacy by design as the default. Health data, security and infrastructure changes get stronger controls; low-risk, easily verified work runs hands-off.
• Lead AI-native engineering by example

Languages