Data pipeline sentinel
Bad data keeps reaching your dashboards quietly — a schema change upstream, a null spike nobody caught — and the data team finds out from an executive. Design a sentinel agent that watches your pipelines for quality problems. Who does it page, when, and how do you stop it from becoming the alert channel everyone mutes? Walk me through the rollout that earns the team's trust, and the pitch that gets it funded.
route_quality_alerts(severities: list[str], dataset_tiers: list[int], weekly_page_budget: int) → list[str][["critical","warning","critical"],[1,1,2],5]out["page","digest","digest"][["critical","critical","critical"],[1,1,1],2]out["page","page","digest"][[],[],3]out[]Turn the fuzzy goal into a bounded agent task. Say what the agent reads, what it may do on its own versus draft for a human, and what it must never touch. Then earn the trust: how you’d verify it before rollout, and the pitch (the problem, what you built, why it matters).
Agent build: design the agent in plain language (or narrate it). The coach grades the decomposition, the guardrails, the verification plan, and the pitch.