Google runs 5‑day AI agents course
Google is circulating a hands‑on five‑day course on AI agents covering tools, MCP, context engineering and production scaling — aimed at rapidly up‑skilling engineers on agent architectures. Expect this to become a common baseline for teams building production agents. (x.com)
Google ran the free five‑day “AI Agents Intensive” in partnership with Kaggle from November 10–14, 2025. (blog.google)) The published syllabus lists Day 1: agent architectures, Day 2: agent tools and MCP interoperability, Day 3: context engineering and memory, Day 4: agent quality/observability and evaluation, and Day 5: prototype‑to‑production including multi‑agent scaling and the Agent‑to‑Agent (A2A) protocol. (blog.google)) Instructional format combined hands‑on codelabs, live Discord discussions and YouTube livestreams, plus a capstone project with prizes and community showcase. (blog.google)) Course artifacts were made available as a Kaggle self‑paced guide and in community GitHub repositories that include notes, code and capstone examples. (kaggle.com)) A five‑slide executive brief that mirrors the course days gives a direct narrative: Slide 1 “Architecture & Strategy” (Day 1), Slide 2 “Tooling & MCP Integration” (Day 2), Slide 3 “Context & Memory Plan” (Day 3), Slide 4 “Quality, Observability & SLOs” (Day 4), Slide 5 “Prototype‑to‑Production, Scaling & A2A” (Day 5). (blog.google)) Deck artifacts aligned to those slides should match course deliverables: an architecture diagram and module map (Day 1), an MCP integration matrix listing APIs and auth scopes (Day 2), a memory budget and session/state strategy (Day 3), concrete SLOs plus logging/tracing dashboards (Day 4), and a deployment runbook with scaling patterns and A2A interoperability points for production rollout (Day 5). (blog.google))
