Overview

This course teaches the skills to design and implement effective multi-agent workflows using the Google ADK framework and Vertex AI Gemini. Starting with foundational architecture patterns, you'll progress through implementation, orchestration (sequential and parallel), custom routing logic, explicit state management, and distributed A2A communication. You'll learn to integrate external databases, implement multi-agent RAG with vector search, and build microservices-based architectures. The final project involves building a distributed banking system with multiple specialized agents communicating via A2A protocol.

Syllabus

Introduction to Multi-Agent Systems with ADK

Get to know your course instructors, set up GCP resources, and get an overview of the course.

Designing Multi-Agent Architecture

Explain the core components of multi-agent systems and how to design their high-level architecture.

Designing Multi-Agent Architectures with Diagramming Tools

Design multi-agent systems using the Orchestrator–Worker pattern and formalize agent capabilities with A2A Agent Cards to enable standardized service communication.

Multi-Agent Architecture with Python

Develop a multi-agent system by coding the designed architecture and connecting agents with well-defined interfaces.

Implementing Multi-Agent Architectures with ADK

Implement multi-agent architectures with ADK by creating specialized sub-agents with focused tools, orchestrators that route intent-based tasks, and structured delegation workflows.

Orchestrating Agent Activities

Apply orchestration techniques to coordinate multiple agent actions and achieve complex workflows.

Implementing Agent Orchestration with ADK

Build complex business processes using SequentialAgent for ordered pipelines and ParallelAgent for concurrent execution, optimizing workflows through intelligent task dependency management.

Routing and Data Flow in Agentic Systems

Configure routing mechanisms to manage data flow among agents in multi-agent systems.

Implementing Data Routing with ADK

Implement deterministic business rule routing with CustomAgent to inspect events and programmatically enforce compliance, discounts, and strict logic without LLM uncertainty.

State Management in Multi-Agent Systems

Evaluate methods for tracking and updating agent state across multi-turn interactions.

Implementing Advanced State Management with ADK

Master explicit session state management using ToolContext and InvocationContext to persist complex data across multi-turn conversations with cloud-based Vertex AI Agent Engine.

Multi-Agent Orchestration and State Coordination

Develop a coordinated multi-agent system that synchronizes states for coherent task execution.

Implementing Advanced State Management with ADK and A2A

Build distributed multi-agent microservices using RemoteA2aAgent for A2A communication, Agent Cards for service discovery, and MCP Database Toolbox for shared MySQL state across independent services.

Multi-Agent Retrieval Augmented Generation

Extend RAG to multiple cooperating agents, each specialized in certain retrieval tasks.

Implementing Multi-Agent RAG with ADK and Vertex AI Search

Implement enterprise multi-agent RAG systems that combine Vertex AI Search for unstructured docs with SQL databases for structured data, enabling hybrid knowledge sharing across agents.

Project: Multi-Agent Banking System

Build a secure multi-agent banking prototype where separate agents handle deposits, loans, and info, share limited data via A2A, and prove safe AI use without real customer data.