
Give Your Agent a Brain: ADK, MCP, & Persistent Memory
Stateless chat is dead. Join this code lab to build intelligent, personalized systems using Google ADK, external tools via MCP, and persistent memory.
About
We are moving past basic wrappers and focusing on developer intelligence. Join us for an exclusive, hands-on engineering sprint where we will break down the architecture of production-grade AI systems. You will learn how to build a robust intelligence layer using Google’s Agent Development Kit (ADK), seamlessly connect it to external databases using the Model Context Protocol (MCP), and give it long-term memory. If you are ready to build systems that actually execute complex tasks and remember who you are, this is where you need to be. What We Are Building During this intensive lab, we will build a stateful, customized agent from scratch. You will learn how to: Master Google ADK: Set up the core intelligence layer using LlmAgent and the Gemini API to drive dynamic reasoning. Integrate MCP Servers: Use the open-source Model Context Protocol to standardize how your system connects to external APIs, databases, and filesystems without writing brittle, custom code. Implement Persistent Memory: Move beyond the standard context window. We will explore lightweight memory structures (like SimpleMem or Mem0) and vector search capabilities (like Qdrant) so your system can recall past interactions, store preferences, and maintain deep context across multiple days. Execute the Handshake: Watch your architecture dynamically discover and load external capabilities through the MCP client-server model. Who Should Attend? This is a technical, code-first environment designed for builders: Software Engineers and Full-Stack Developers AI Researchers and ML Practitioners Founders building AI-native products and developer tools Prerequisites (What to bring) To code along with the room, please ensure you have: A laptop and charger. Your preferred IDE (VS Code, Cursor, etc.). Basic proficiency in Python or TypeScript. An active Google AI Studio account for API keys. Familiarity with standard APIs and basic asynchronous programming concepts is highly recommended. Event Schedule 11:00 AM: Doors Open, Coffee, & Environment Setup 11:30 AM: The Architecture: ADK Fundamentals & The Intelligence Layer 12:30 PM: Lunch Break 🥪 1:00 PM: Code Lab 1: Implementing Persistent Memory & Vector Storage 2:15 PM: Code Lab 2: Connecting the Real World via MCP Servers 3:15 PM: Live Demos: Testing Your Stateful Architecture 3:45 PM: Q&A & Wrap-up 4:00 PM: Event Concludes
This hackathon has ended