Build AI Agents with Google ADK

Name: Build AI Agents with Google ADK
Rating: 4.7 (8 reviews)

Hands-on AI Agent development with Google ADK, Gemini models, and Vertex AI ,multi-agent systems

Created byRiad Almadani • 70,000+ Students

Last updated 6/2026

English

What you'll learn

Build real-world AI Agents using Google ADK from scratch
Design custom tools and integrate built-in tools like Google Search & Gemini
Implement short-term and long-term memory in AI agents
Connect agents to Vertex AI for scalable memory systems
Create production-ready AI workflows with state and session management
Build projects like a Hello Agent and Real-Time Weather Agent
Understand agent architecture (Agents, Tools, Workflows) deeply
Design intelligent systems that think, remember, and act autonomously

Course content

5 sections • 24 lectures • 3h 27m total length

Welcome and Course Overview3:05
Learn to build ai agents with Google's ADK, starting from basics and progressing to memory, tools, web search, and deployment in the cloud.
The Rise of AI Agents4:23
Explore why AI agents matter and how Google's ADK enables developers to build memory-enabled, multi-step agents that plan, act, use real tools and APIs, and deploy to Cloud Run.
ADK Architecture – Agents, Tools, and Workflows5:56
Explore how ADK uses agents, tools, and workflows to coordinate tasks, with agents as brains, tools as hands, and workflows as the plan, including sequential, parallel, and smart decision types.
Setting Up Your Environment4:04
Set up your environment by creating a Google Cloud project, generating an ADK-course API key via Google AI Studio, and installing Python and VS Code to run Google ADK locally.
Hands-on Project – Building Your First “Hello Agent”7:07
Create your first ai agent HelloAgent with adk, configure Gemini 2.5 flash via Google AI, and run agent.py to chat through the web interface.

Creating Custom Tools21:27
Create custom tools for an agent with the ADK, building a calculator tool described by a docstring, test it, and see the agent use it to solve math prompts.
Using Built-in Tools & Gemini Capabilities (Google Search)3:52
Learn to empower AI agents with built-in tools in Google ADK, including Gemini model integration, using Google search to perform live internet queries and run a search-enabled agent.
Hands-on Project – Building “Real Time Weather Agent”7:23
Build a weather agent that fetches live data from the internet using the requests library. Connect a weather tool to your ADK agent to query cities like New York.

Understanding Agent State and Sessions5:37
Learn how state and sessions enable memory in adk, using implicit state and automatic memory to remember conversation history within a temporary session.
Short Term Memory3:02
Explore how short-term memory works in ADK through a context window that feeds the conversation to Gemini, enabling the agent to remember names and follow-ups within a session.
Long Term Memory3:23
Save your AI agent's memory for long-term use by enabling the adk run --save-session flag. Resume from a session file like user1.session.json to recall previous conversations.
Long Term Memory With Vertex AI6:14
Move memory from a local JSON file to the cloud with Vertex AI Memory Bank Service, ensuring persistent, scalable memory and smart retrieval for many users.
Long Term Memory With Vertex AI Part17:00
learn to use Vertex AI memory bank service with Google ADK to build agents in AI Studio, following system instructions and plain Python functions for reliable, copy-paste ready code.
Long Term Memory With Vertex AI Part26:25
Learn to build and adapt a complex agent with Vertex AI memory bank for long-term memory, run via ADK web, using set user name and memory search tools.
Long Term Memory With Vertex AI Part312:19
Learn to implement the set username tool in a Vertex AI powered agent, storing user names in session memory to personalize learning and track progress across sessions.
Long Term Memory With Vertex AI Part49:49
Learn to implement long-term memory for a tutoring agent with Vertex AI memory bank, using set user name and asynchronous memory tools to personalize learning across sessions.
Long Term Memory With Vertex AI Part55:24
Long Term Memory With Vertex AI Part615:53
Long Term Memory With Vertex AI Part719:02
Long Term Memory With Vertex AI Part816:54

Requirements

Basic Python knowledge
No prior AI agent experience required

Description

If you’re new to generative AI and curious about building real AI tools, this course will take you from zero to job-ready in Google’s Agent Development Kit (ADK). You’ll learn to build intelligent AI agents powered by Google’s Gemini models and Vertex AI platform. We start with the basics – setting up your Python environment, installing ADK, and creating your first agent in minutes – and then add complexity step by step. Through hands-on, project-driven lessons, you will create custom “tools” (APIs) so your agents can do things like book flights, manage calendars, or analyze data. Along the way you’ll master key concepts like agent state and memory (persistent sessions and databases), structured outputs.

Later in the course, you’ll see how Google’s Vertex AI makes it easy to train models, perform Retrieval-Augmented Generation (RAG) with vector search, and deploy agents to production. You’ll build multi-agent systems where specialized agents collaborate on complex tasks, and even implement “agent-as-a-tool” scenarios where agents call other agents. Throughout, the focus is on real-world readiness: we use practical examples and code, emphasize debugging and testing, and cover deployment on Google Cloud.

By the end of this course, you won’t just have theory – you’ll have a portfolio of AI agent projects and the skills to apply them. Instructors promise you’ll “build, deploy, and scale production-ready AI agents” using Google ADK. This beginner-friendly course requires only basic Python; no prior AI framework experience is needed. You’ll finish able to design, debug, and launch autonomous AI agents that use Gemini, manage their own memory, and solve realistic problems – exactly the job-ready skills employers are looking for.

Who this course is for:

Python developers who want to build real AI agents instead of simple chatbots
Developers interested in Google ADK and Gemini
AI enthusiasts who want to understand how intelligent systems work intern
Software engineers looking to add AI agent architecture to their skillset
Students who want hands-on projects, not just theory
Anyone who wants to move from “prompt engineering” to real AI system design

Build AI Agents with Google ADK

What you'll learn

Explore related topics

Course content

ADK Introduction5 lectures • 25min

Core ADK Concepts – Agents and Tools3 lectures • 33min

State, Memory, and Context Management12 lectures • 1hr 51min

Multi-Agent Systems2 lectures • 16min

Deploy Your Agent2 lectures • 23min

Requirements

Description

Who this course is for: