AI agents are great at generating answers, but what happens when they forget everything after the conversation ends?

Join the Dublin MongoDB User Group for an evening of hands-on building, practical AI engineering, and a bit of friendly competition. We’ll explore how to give AI agents memory, how to retrieve the right information at the right time, and how small retrieval design choices can make a big difference.

!!Bring your laptop!! You’ll build a memory-augmented AI agent with MongoDB then put your retrieval skills to the test in a fun skills challenge with leaderboard and prizes.

And yes, you can walk away with a Memory for AI Applications Skill Badge to share with your network.

This event is part of Dublin AI Week, taking place from 2 to 9 October 2026. Check out the other AI events happening across Dublin and build your own week of learning, building, and networking.

What we’ll be covering ----------------------------

The “Why”: Why AI agents need memory: Learn how short-term and long-term memory help agents maintain context across conversations, remember useful information, and support longer-running tasks.

“Build”: Agent Memory hands-on workshop: Build a memory-augmented coding agent and explore:

  • Core memory types for agentic systems
  • How to create, read, update, and delete memories
  • Storage patterns for reliable retrieval
  • Combining vector search with metadata, recency, and other signals
  • When to create, update, or prune memories

The workshop uses Jupyter Notebooks and a pre-configured development environment. You’ll work from templates and replace the pseudocode with your own Python code.

The “Challenge”: Chunking, Retrieval & Reranking

In this practical challenge, you’ll experiment with chunking strategies, retrieval methods, and reranking in a GUI to improve the quality of search and answers. We’ll explain the benefits, trade-offs, and best practices behind each approach, then let you configure the retrieval pipeline and submit your results to receive a score.

Compare your approach with other attendees on a leaderboard, learn from the results, and compete for prizes.

Agenda

  • Networking, pizza, and introductions 🍕
  • Agent Memory hands-on workshop
  • Retrieval challenge: chunking, retrieval, and reranking
  • Skill badge, open discussion, and more networking

Full timings will be shared once the agenda is finalised.

What to bring

  • A laptop — this is a hands-on event
  • Access to the workshop environment through Instruqt
  • A work or personal account for the workshop environment
  • Optional: a quick look at the Jupyter Notebook guide

The AI Labs setup guide explains the available development environment options. You don’t need to be an AI or MongoDB expert, but you should be comfortable reading and writing code.

Want to speak / share your experiences with MongoDB? Have a MongoDB, AI, or developer-focused topic you’d like to share? Submit it through the MUG speaker form. We may be able to include it in this event or a future MUG.

See you there! 🎯

  • agentic ai
  • ai
  • ml