AI Agents with Google’s ADK Bootcamp

Get hands-on with Google's Agent Development Kit (ADK) and learn how to build powerful AI agents that reason, act, and collaborate.

Overview

This immersive two-day live course takes you from zero to launch-ready agents, walking you through the core tools, techniques, and protocols behind production-grade agent development. You'll connect Gemini and other LLMs, define custom tools, implement memory, master patterns like RAG, and explore security, deployment, and evaluation. If you're looking to build smarter software that talks, thinks, and helps—this is your launchpad.

Train up your teams with private group training

Have a group of 5 or more students? Cprime also provides specialist private training with exclusive discounts for tailored, high-impact learning.

Full course details

Course outline

Part1: Intro to Agents & ADK

  • What are AI agents and why now?
  • ADK core concepts (agent, tool, memory, workflow)
  • Demo: adk init structure walkthrough
  • Hands-on: Create your first agent
  • Discussion: Agent use cases
  • Q&A

Part 2: Models and Tools

  • Connecting to Gemini via Vertex AI
  • Using open models via LiteLLM
  • Defining Python tools and schemas
  • Demo: Agents in action with Gemini + tool use
  • Hands-on: Create a Python tool and link it to your agent
  • Q&A

Part 3: Memory and Multimodality

  • Conceptual flow and memory persistence
  • Types of memory: buffer, summary
  • Implementing memory in ADK
  • Hands-on: Add memory and interact using adk web
  • Discussion: What agents should remember
  • Q&A

Part 4: Agentic Design Patterns

  • RAG and beyond: why patterns matter
  • Demo: Search + LLM retrieval
  • Hands-on: Implement a design pattern of choice
  • Group share: favorite patterns and use cases
  • Q&A

Part 5: Building the Full Agent

  • Review core building blocks
  • Hands-on: Create a "Daily Briefing Agent"
  • Discussion: Day 1 highlights and blockers
  • Q&A

Part 6: MCP and Agent-to-Agent Protocols

  • MCP and A2A intro
  • Use cases for multi-agent systems
  • Demo: Make your agent A2A-compatible
  • Hands-on: Swap tools, connect to MCP server
  • Discussion: Multi-agent  architectures
  • Q&A

Part 7: Agent Evaluation and AgentOps

  • Evaluation strategies for agents
  • AgentOps overview: lifecycle, CI/CD, logging, testing
  • Demo: adk run -v, agent engine, deployment paths
  • Hands-on: Analyze logs, prep for deployment
  • Q&A

Part 8: Agent Security & Wrap-Up

  • Security risks: prompt injection, tool misuse, data leakage
  • ADK mitigation strategies
  • Project: Identify one or two risks that apply directly to the agent you’ve conceptualized or built; explore the different capabilities in ADK to secure your agent.
  • Hands-on: Identify and address risks in your agent
  • Final Q&A + Certification Badge Instructions
Audience / prerequisites

You will need:

  • A computer with Python installed and working CLI access
  • Code editor (e.g., VS Code) and browser access
  • Google Cloud Platform account (with billing enabled/free credits)
  • Basic knowledge of Python, APIs, and LLM principles
  • Comfort using the terminal and following setup instructions

This course is great for

  • Software engineers and backend developers building intelligent systems.
  • AI/ML engineers seeking production-ready agent frameworks.
  • Solution architects evaluating agent stacks and infrastructure options.
  • Technical leads and product teams prototyping AI-driven automation

 

In this class you will learn how to
  • Understand the agent architecture and core ADK workflow.
  • Build and test agents using Gemini, open models, and Python tools.
  • Apply patterns like RAG and memory buffers in real-world tasks.
  • Integrate agents with Model Context Protocol (MCP) and A2A.
  • Evaluate, monitor, and deploy agents with confidence using AgentOps principles.
  • Address safety and security risks using best-practice guidelines.
Cancellation Policy

A full refund will be issued for class cancellations made at least 10 business days before the course begins. Payment is nonrefundable for cancellations or reschedules made within 10 business days from the course start date and for no-shows (students who do not attend class). For reschedules made within 10 business days from the course start date, students will be responsible for a $200 administrative fee.

Request Private Group Training