Belgian Java User Group

Finding Java Memory Leaks, AI Agentic Setup and Fraude Detection

Tue, Mar 24 · 6:00 PM CET

Tour & Taxis, Picardstraat 7, Brussels, BE

We are going back to Brussels, which means we'll once again be teaming up with the Brussels Java User Group <3.
We'd like to thank Adesso for graciously hosting!

This edition features 3 talks!

• Finding Java Memory Leaks with Heap Dumps (JDK 25) by Badr Kacimi
• Vibe Coding a Home Server on the Raspbery Pi, What AI-Native Development Actually Looks Like by Wim De Troyer
• Fraud detection on card transactions, Supervised learning with extreme imbalanced data using Python by Franck Benault

Come join us for this next event!

Schedule
- 18:00: Doors open + Food and Drinks
- 18:45: Intro and BeJUG updates
- 18:50: Talk 1 - Fraud Detection
- 19:05: Talk 2 - Vibe Coding a home server on the raspberry pi.
- 19:30: Little break
- 19:45: Talk 3 - Finding Java Memory Leaks with Heap Dumps (JDK 25)
- 20:30: Mingling + Drinks

Topics

Finding Java Memory Leaks with Heap Dumps (JDK 25)
This talk will demonstrates how to identify, analyze, and fix Java memory leaks using heap dumps and practical JVM diagnostic tools in JDK 25.

About Badr
Badr Kacimi is a Senior Software Engineer at ABN AMRO Bank and GenAI Whisperer at Blue Harvest Belgium.

With a strong Java background in large-scale banking and booking platforms, he now builds financial risk models in Python.

Vibe Coding a Home Server on the Raspbery Pi: What AI-Native Development Actually Looks Like

It's been over 3 years since the release of ChatGPT and the innovations in the AI space still do not seem to be slowing down.
On the contrary, agentic tooling like Claude Code and Codex have since launched, kept improving, and according to some have now recently reached an inflection point where they can handle long running coding tasks (mostly) unsupervised.

Let's find out together if those claims hold up and what that new AI native development looks like.
We'll do this by showcasing a very personal case study.
Over the past few months, I've developed a home-server setup on my raspberry pi with tons of weird and quirky apps all developped with varying degrees of AI assistance.

Join me to find out how that went, how much time I saved (or wasted!) and my biggest take-aways and future plans.

About Wim
Wim is a freelance Software Engineer from Belgium with around 7 years of experience. He focuses on backend web development, using the Spring Framework on a daily basis.
He's a firm believer in learning something new everyday (not just about java or spring!) and doing so in public is always the greatest motivator.
So he blogs on wimdetroyer.com, and talks occasionally in front of an audience when he can muster up the courage.

Fraud detection on card transactions, Supervised learning with extreme imbalanced data using Python

I will explain:

• The context and the goal of this work (supervised learning)
• The challenges related to the data (availability and preparation)
• The strategies to apply when dealing with severe class imbalance

About Frank
Research and development engineer in computer science, Java developer, passionate about physics, astronomy, and cosmology.