Join us for a MongoDB User Group in Sydney 🚀

Come connect with fellow developers, tech enthusiasts, and the curious-minded for an evening exploring the best practices and real-world use cases powering modern apps with MongoDB. Whether you're building your first project or scaling to millions of users, you'll walk away with practical insights, fresh ideas, and a great group of people to swap notes with.

Grab a drink, meet the local MongoDB community, and see what's possible when you build with the right database. See you there!

Agenda

17.00: Arrival & Networking18.00: Welcome18.10: MongoDB Atlas Search under the Hood 🔍18.40: Your LLM Is Burning Cash. MongoDB Might Be the Fire Extinguisher 💰🧯19.10: Pizza & Networking20.00: Event Close

Session Details

MongoDB Atlas Search under the Hood 🔍Luke Thompson, Chief Architect, Clarifresh

Let's explore the newly open-sourced service MongoT, a Java sidecar application which runs alongside the C++ MongoD service providing additional search capabilities. We'll take a quick look around the codebase, spin it up and show how real Atlas Search queries turn into Lucene index searches. Bio: Luke Thompson is a software architect and passionate coder building software by day and by night. He applies his talents, such as they are, to make things as good as they can be.

Your LLM Is Burning Cash. MongoDB Might Be the Fire Extinguisher 💰🧯Jinu VM, Advisory Solutions Architect, MongoDB

Every token you send to an LLM costs money and most teams are sending way more than they need to. Bloated context windows, redundant retrieval, and poorly managed memory are quietly driving up inference costs and slowing down response times, all while making your AI less accurate, not more.In this session, we'll dig into how smart context engineering and memory management can put out the fire. You'll learn how MongoDB can serve as the backbone for leaner, smarter AI applications. From efficient vector search and retrieval strategies that cut down on unnecessary context, to persistent memory architectures that let your LLM remember what matters without re-sending it every time.What you'll walk away with:

  • Why bigger context windows aren't always better (and often cost you accuracy and cash)

  • Practical patterns for structuring memory: short-term, long-term, and episodic

  • How vector search and retrieval-augmented generation (RAG) can trim your token spend without sacrificing quality

  • Real-world architecture patterns for building cost-efficient, context-aware AI applications on MongoDB

Whether you're building agents, chatbots, or RAG pipelines, this session will help you rethink how you manage context so your LLM budget stops going up in smoke.