Staff AI Engineer, Agent Platform (AI Incident Response)

A venture-backed company has built an AI agent that investigates production incidents on its own, correlating telemetry such as logs, metrics and traces with change records so on-call engineers receive an actionable root cause. It is hiring a senior individual contributor to own the platform that runs those investigations and to set its technical direction.

Platform ownership

The next platform architecture: how the agent accumulates knowledge across investigations and extends its reasoning over many data sources while keeping customer tenants isolated and answers grounded

Reliability of the runtime that executes investigations under time pressure against many integrations, including checkpointing, recovery, graceful degradation, time and cost budgets and failure telemetry

Output backed by evidence, so confidence reflects retrieved data and recommendations match each customer’s actual tooling

The evaluation system that gates every release, built on real incidents, with regression checks and quality measures

Technical leadership: design reviews, platform standards and mentoring of other AI engineers

Expertise required

Production experience with ML or LLM systems spanning data, models or agents, deployment and evaluation, where quality was measured

Strong production engineering fundamentals and ownership of reliability

Evaluation-driven development: eval sets you designed, releases gated on them and metrics you improved

Strength in orchestrating agents or in persistent memory and context, with working knowledge of the other

Bonus: distributed systems, observability or AIOps exposure, multi-tenant data isolation, grounding controls in regulated settings

On-site five days a week on the San Francisco Peninsula (Redwood City area). Visa transfers and new sponsorships are both considered.