Build with AI: Building Personalized Agents with ADK, MCP, and Memory Bank

Hosted by Orca AI Club & 8 others
Registration
Past Event
Welcome! To join the event, please register below.
About Event

​Learn to Build. Learn by Doing.

​Join the #BuildwithAI Movement.

​Stop consuming and start creating. In this hands-on technical lab, we are moving beyond simple prompting. We are guiding you through building an Autonomous AI Agent that doesn't just "chat" but utilizes a Memory Bank and the Model Context Protocol (MCP) to interact with the real world and remember complex user context.

​What We Are Building:

  • ​A Persistent AI Agent: A functional system capable of long-term memory and personalized reasoning.

  • ​Built using the Google ADK & MCP: Leveraging the latest Model Context Protocol to bridge LLMs with local data and external tools.

  • ​Memory Integration: Implementing a "Memory Bank" logic to handle high-entropy data without losing task focus.

​Session Agenda:

  • ​Agentic Foundations: Introduction to the Agent Development Kit (ADK) and the paradigm shift from Chatbots to Agents.

  • ​Protocol Setup: Configuring the Model Context Protocol (MCP) to allow Gemini to "see" your local environment and files.

  • ​Architecture Build: Constructing the "Memory Bank" using Python to store and retrieve persistent user personas.

  • ​Logic Development: Coding the decision-making loop—enabling your agent to choose the right tool for the right task.

  • ​The Enigma Stress-Test: Testing the agent’s selective attention by introducing distractors and verifying its attentional persistence.

  • ​Deployment & Scaling: Discussion on moving from local Python scripts to production-ready agentic workflows.

  • ​Q&A

​Important: Event Prerequisites

  • ​Required Hardware: A laptop and charger are mandatory. This is a high-intensity coding lab unsuitable for mobile devices or tablets.

  • ​Environment Setup: Please have Python 3.10+ and a code editor (VS Code recommended) pre-installed. We will provide the API keys and MCP configurations on-site.

  • ​Skill Level: Intermediate. Familiarity with Python and basic API concepts is expected, as we will be diving into agentic architectures and memory state management.

​Host Community:

​This event is run by Orca AI Club in collaboration with Amrita School of AI & Data Science, supported by Build with AI.

Location
Amrita Theatre Complex
C9G5+43R, RPS City, Sector 88, Faridabad, Haryana 121014, India