Google Agent Development Kit (ADK) Crash Course

In this short crash course, we take you on a fun, hands-on and pragmatic journey to learn how to build Multi-Agent Systems using Google's Agent Development Kit (ADK). You'll start building your first AI Agent within minutes. Every section is recorded in a bite-sized manner and straight to the point as I don’t want to waste your time (and most certainly mine) on the content you don't need.

In this course, we will cover:

1. Introduction & Setup — Installing ADK, exploring the project, running your first agent.

2. Built-In & Custom Tools — Understanding available tools and creating your own.

3. Model Connections — Connecting to different LLM models (LiteLLM).

4. Structured Outputs — Working with structured responses in agents.

5. Sessions, State, and Runners — Core components and how agent state is managed.

6. Persistent Sessions & Memory — Building reminder agents, storage, and database usage.

7. Multi-Agent Systems Overview — How agents work together and share state.

8. Agents as Tools & Deep Code Walkthrough — Using agents as callable tools and exploring the code.

9. Trading System Agents — Root, Policy, Trade, Stock Research, and Order Agents.

10. Callbacks System — Before/after agent, model, and tool callbacks.

11. Sequential & Parallel Agents — Workflow agents, running sequential and parallel pipelines.

12. Looping Agents — Creating and running looping/iterative agents.

The goal of this course is to teach you ADK development in a manageable way without overwhelming you. We focus only on the essentials and cover the material in a hands-on practice manner for you to code along.

Working Through This Course

This course is purposely broken down into short sections where the development process of each section will center on different essential topics. The course a practical hands on approach to learning through practice. You learn best when you code along with the examples.