Meet Your Instructor
Maurizio Ipsale
Google Cloud Authorized Trainer · Google Developer Expert (AI & Cloud)
This course exists because of a simple frustration: most ADK material either stays too shallow to build anything real, or drops you into production concerns before you've built your first agent. As a Google Cloud Authorized Trainer and GDE, I built the course I wished existed — 40 modules, zero shortcuts, every concept backed by a lab you actually have to solve.
How this course works
You don't learn to build agents by reading about them.
Every module follows the same loop — and the "Hidden Solution" is exactly that: hidden. You have to actually get stuck first.
01 · README.md
Theory
The concept, explained once, with the exact ADK 2.0 API you'll use next.
→02 · lab.md
Challenge Lab
A working skeleton with real TODOs. No copy-paste — you write the logic.
03 · lab-solution.md
Solution
Base64-hidden on purpose. Decode it only once you've actually tried.
Curriculum
Seven parts, one continuous graph
Each part builds on the state and skills of the one before it — same as an ADK workflow, nothing here runs out of order.
The absolute basics of AI agents and the ADK — environment setup through your first multimodal agent.
Custom function tools, OpenAPI, built-in grounding, and third-party integrations.
Static, structured, dynamic, and cyclic orchestration — up through distributed A2A graphs.
State, memory, artifacts, evaluation, observability, and callback-based guardrails.
MCP clients and servers, plus wiring a custom streaming front-end to your agents.
Cloud Run, GKE, and enterprise-grade Gemini Enterprise Agent Platform deployment.
Plugins, agent skills, and a full enterprise incident-response capstone with AgentOps.