Google Agent Development Kit (ADK)-Build & Deploy AI Agents

A hands-on, code-first course covering ADK architecture, tools, memory, callbacks, guardrails, and deployment

What you’ll learn
Understand the building blocks of LLMs and how they power intelligent agents
Set up and scaffold your first AI agent using Google’s Agent Development Kit (ADK)
Run agents through the ADK Web UI, CLI, and REST API
Master ADK’s architecture including the Runner, Sessions, State, and the Event Loop
Connect agents to Gemini, GPT, Ollama, and third-party models through Vertex AI and LiteLLM
Build and integrate Function Tools, Built-in Tools, OpenAPI Tools, and MCP Tools
Use LangChain and CrewAI tools within ADK agents
Implement session-based memory, shared state, and long-term memory for persistent agents
Add callbacks and guardrails to control agent behavior before and after model and tool execution
Prepare agents for production and deploy them to Google Cloud Run
Deploy agents to Vertex AI Agent Engine for managed, scalable hosting

Requirements
No prior experience with AI agents or agent frameworks is needed
Basic Python programming skills (comfortable with functions, dictionaries, and installing packages)
A Google Cloud account with billing enabled (free tier is sufficient for most labs)
A Google AI Studio API key (free to obtain)

Description
Every major framework gives you a way to build AI agents. But most courses teach you to build toy demos that never leave your laptop.

This course is different.

You will go from your very first agent to deploying a production-ready, guardrailed system on Google Cloud, entirely within Google’s Agent Development Kit (ADK).

ADK is not another wrapper around an LLM. It is a purpose-built agent framework with first-class support for tool integration, session management, shared state, long-term memory, lifecycle callbacks, and native Google Cloud deployment. Unlike courses that hop between five frameworks in a weekend, this one goes deep on one.

The course spans 11 carefully sequenced modules. Foundations (Modules 1 through 5) take you from zero to running agents with the Web UI, CLI, and API. Pillars (Modules 6 through 10) go deep into architecture, models, tools, memory, callbacks, and guardrails. Deployment (Module 11) gets your agent running on Cloud Run and Vertex AI Agent Engine.

Every module follows a consistent rhythm. Concept slides, live demos, hands-on labs, and a quiz.

This is a code-first, hands-on course. Not for no-code builders or Python beginners. For developers serious about building AI agents that actually ship.

Enroll and let’s build.

Who this course is for
Python developers who want hands-on experience building AI agents with Google ADK
Cloud engineers and architects evaluating agentic AI solutions on Google Cloud
Software engineers looking to move from chatbot prototypes to production-grade agent systems
AI practitioners who want deep, focused training on a single framework rather than a shallow multi-framework survey
Anyone who wants to deploy real agents to Cloud Run or Vertex AI Agent Engine