Google Agent Labs: Building Multi-Agent Systems with ADK
The Launch: AI Agent Builder Series 2026
AI House Bangalore × Google for Developers
Huge milestones call for high-execution build sessions. AI House Bangalore has officially partnered with Google for Developers to launch the AI Agent Builder Series 2026 India's most ambitious Agentic AI builder program spanning 45 days, 10+ cities, and over 10,000 builders, culminating in a Grand Finale at the Google Office, Bengaluru.
This lab session marks the official kickoff event for the entire national series.
Google Cloud recently introduced the Agent Development Kit (ADK) at NEXT 2025, a new open-source framework purpose-built to help you design, deploy, and evaluate intelligent multi-agent systems with ease. Whether you’re looking to orchestrate agents, integrate tools, or deploy production-grade agentic applications, ADK provides everything you need in one developer-friendly toolkit.
According to the official blog, the Agent Development Kit is:
a powerful, open-source framework designed to simplify the development of production-ready AI agents and multi-agent systems with flexible tooling, rich integrations, and deployment options built for scale.
Here at AI House Bangalore, we prioritize execution over consumption. In this lab session, we’ll go beyond the basics and get hands-on. You will write code to explore what it really takes to transition from a simple prompt to a fully deployed, collaborative agent workflow using Google's ADK.
What We’ll Explore
What does it really mean to build multi-agent systems with ADK?
How do agents delegate tasks, stream audio/video, and interact with tools like LangChain, custom APIs, or Vertex AI models?
How do you test execution trajectories, manage state, and deploy these agents into real-world products?
What You Can Do with ADK
Build modular multi-agent systems with clear hierarchical delegation and orchestration.
Integrate with any model via Vertex AI Model Garden or LiteLLM.
Use tools and libraries including custom Python functions, LangChain tools, and Anthropic's MCP (Model Context Protocol) servers.
Enable bidirectional streaming for rich, natural multimodal interaction natively out-of-the-box.
Evaluate performance before deployment using built-in test frameworks against predefined datasets.
Deploy anywhere, including standard container runtimes or Google Cloud's Vertex AI Agent Engine.
Lab Schedule
Environment Setup & Architecture Breakdown
02:00 PM
Doors open. We'll set up the Python virtual environment, install google-adk, configure your Google AI Studio or Vertex AI keys, and break down the design philosophy of the framework.
Lab 1: The Single Agent Foundation
02:30 PM
Building your first autonomous agent using the Agent class. We'll write structured system instructions, equip it with native tools (like Google Search), and run execution loops locally via the ADK CLI.
Lab 2: Multi-Agent Graphs & Workflows
03:30 PM
Scaling up to collaborative systems. You'll use the Workflow class to define a graph-based multi-agent architecture where a coordinator agent dynamically delegates tasks to specialized sub-agents.
Lab 3: Advanced Tools & Model Context Protocol (MCP)
04:30 PM
Giving your agents arms and legs. We'll explore connecting your multi-agent mesh to external data silos, secure runtimes, and local developer environments using custom tools and MCP servers.
Evaluation & Production Deployment
05:15 PM
Testing execution paths using the built-in AgentEvaluator. We'll walk through debugging agent trajectories and detail how to containerize and push your system to the Vertex AI Agent Engine.
Showcase & Peer Networking
05:45 PM
Demo your working agent workflows, gather feedback from senior AI architects in the room, and sync up with fellow builders.
What You’ll Learn
How to build and structure production agent applications using ADK’s Pythonic, code-first framework.
How to give agents real-world capabilities by connecting them to external APIs and databases via MCP.
How to construct deterministic logic and adaptive AI reasoning using Graph Workflows.
Who Should Attend
AI Engineers who want to build production agentic applications using Google's latest open-source tools.
AI Engineering Leaders who want to evaluate the newest enterprise-grade agent orchestration frameworks.
Hosted By
Google for Developers is Google’s global ecosystem dedicated to helping developers build, ship, and scale next-generation applications. From fundamental infrastructure to cutting-edge AI orchestration like Gemini and the newly launched open-source Agent Development Kit (ADK), Google provides the open tools, managed platforms, and community frameworks that power the developer intelligence ecosystem worldwide.
AIHouse is the community layer powering this ecosystem. Built for founders, researchers, engineers, students, and AI enthusiasts, AI House creates a space where people consistently learn, share, collaborate, and grow together. Through initiatives like 100 Days of AI House, community gatherings, founder conversations, knowledge-sharing sessions, and industry collaborations, AI House has built a thriving network of 15,000+ community members who are actively shaping the future of AI. More than a community, AI House is where meaningful relationships, opportunities, and ideas are created every day.