Linear expert for agent-driven engineering org — design our workspace for parallel AI coding agents

We're a small EdTech startup building a large, spec-heavy platform. Our development is agent-first: an orchestrator decomposes epics into dependency-ordered work packages and dispatches parallel AI coding agents that open real PRs. Linear.APP is the spine — the single source of intent, dispatch, review gates, and audit trail. We came from Jira and we know our Linear setup doesn't yet match how we build. We need someone who has engineered this, not read about it: Linear agent app-users, webhook-driven dispatch, issue models that survive high-frequency agent updates, human review gates inside the workflow, and notification design that doesn't drown four humans. Design the workspace: teams/projects/issue structure to hold work packages and their dependencies Wire the agent side: app-user, webhook subscriptions, delegation flow, idempotency Build a human "understanding review" gate state with SLA + escalation, using stock Linear Notification architecture for 4 named humans with different roles Views/dashboards: founder-readable progress without asking engineers Deliverable: working workspace + a short written runbook our team owns afterward Hard requirement: production experience running multiple coding agents against one large codebase with Linear as the control plane. Personal-productivity agents (email, calendar, scraping) do not qualify. In the interview you'll screen-share the systems you describe.