JCON EUROPE 2026 - Workshops (original) (raw)

A workshop ticket costs €19 (incl. tax).
Please note that workshop tickets are completely separate from conference tickets:

• A conference ticket does _not grant access to any workshop.
• A workshop ticket does _not grant access to the conference.
• Each workshop requires its own ticket. If you want to join multiple workshops, you will need one ticket per workshop.

Workshop Day & Schedule

The workshops will take place on April 20, 2026, at Startplatz (a side building next to the Cinedom).

We will offer two 2-hour workshop slots:

• 13:00 – 15:00: Workshop Slot 1
• 15:00 – 16:00: Break & networking
• 16:00 – 18:00: Workshop Slot 2

This means you can attend up to two workshops on this day, one in each slot, and use the break for networking with other participants.

Note: If you buy a workshop now and see that a workshop is released later in the same timeslot that you would like to attend, please email us at info@jcon.one with your order code and the workshop you would like to switch to.

Retrieval-Augmented Generation (RAG) is one of the key building blocks of modern AI systems—yet most implementations remain abstract or depend heavily on external platforms. In this workshop, you’ll take a different path: building a fully self-hosted RAG system in pure Java, step by step.

Using a consistent demo application, you’ll work with Vaadin, EclipseStore, and jVector—deliberately without Spring—to create a setup that runs locally and keeps sensitive data within your own infrastructure. The result is a practical, production-relevant architecture designed for environments where security, control, and data privacy truly matter.

Guiding you through this is Sven Ruppert, a Java expert with nearly three decades of experience across industries like banking, aerospace, and global organizations such as the UN and World Bank. Now focused on application security, he brings a perspective that goes far beyond implementation—showing not just how to build RAG systems, but how to build them securely and responsibly.

Sven Ruppert

DA Security @ Neosec IT

Building GenAI applications often comes with an unexpected cost: external vector databases that complicate development, increase operational overhead, and break the simplicity of a pure Java stack.

This workshop introduces a different approach.

By combining the Java vector search engine JVector with the Java-native persistence engine EclipseStore, you’ll work with a fully integrated vector storage solution built entirely in Java. No additional infrastructure, no context switching—just a clean, cohesive architecture that supports everything from embedded use cases to distributed GenAI applications.

This approach is shaped by Florian Habermann, CTO and Co-Founder of MicroStream, who has spent over two decades rethinking how Java handles data persistence. His work focuses on eliminating unnecessary complexity and bringing developers back to a more natural, object-oriented way of building systems.

Florian Habermann

CTO & Co-Founder @ MicroStream

If your system feels slow, the usual answer is: add a cache.
Still slow? Add a search server.
Need more? Maybe a graph database… or a vector store.

Sound familiar?

What starts as optimization quickly turns into a patchwork of technologies, each adding more complexity, more DevOps effort, and more cost. This workshop challenges exactly that pattern.

Instead of stacking solutions, you’ll learn how to simplify your data architecture, reduce database pressure, and boost performance at the core—using a clean, Java-centric approach built on open-source technologies. The goal isn’t another tool. It’s fewer moving parts and systems you can control.

Christian Kuemmel brings the perfect perspective for this. As a Senior Project Lead at MicroStream, he works daily with distributed systems, cloud-native setups, and scalable SaaS platforms—seeing firsthand how complexity creeps in, and what it takes to keep systems efficient and manageable.

Christian Kuemmel

Senior Project Lead @ MicroStream

Modernizing a Java application can feel like running a gauntlet: upgrade the JDK, untangle dependencies, refactor legacy code, fix the build, then make it cloud-ready—without breaking everything that currently pays the bills. In this hands-on workshop, Sandra Ahlgrimm (Senior Cloud Advocate at Microsoft) and Bruno Borges (Principal Product Manager at Microsoft) show how to speed up that journey by combining AI assistance (GitHub Copilot) with **deterministic automation (OpenRewrite)**—so you move faster _and keep control.

You’ll work on a real project and leave with something you can use immediately at work or on side projects: how to assess a codebase, create an upgrade plan, apply automated transformations, fix the fallout, and validate results. The key takeaways are safe prompting patterns, repeatable modernization recipes, and a practical process for turning “legacy” into “maintainable and cloud-ready” step by step.

Sandra Ahlgrimm

Senior Cloud Advocate @ Microsoft

Bruno Borges

Principal Product Manager @ Microsoft

Some testing habits are like floppy disks: nostalgic, but not something you want in your daily workflow. With JUnit 6 out, this hands-on workshop helps you actively unlearn legacy patterns (yes, including the old “everything starts with test” era) and move your codebase toward a cleaner, modern testing style. Tim te Beek (Solution Engineering Manager at Moderne) brings the practical migration experience of someone who helps teams refactor at scale—without turning CI into a slot machine.

You’ll work through real upgrade steps (bring your own project if you like): replacing brittle, dated assertion styles such as Hamcrest with AssertJ 4, simplifying test structure, and tightening feedback loops so failures are easier to read and faster to fix. You’ll leave with a repeatable approach you can apply back at work or in side projects—so your tests become boring, stable, and maintainable for the decade ahead.

Tim te Beek

Solution Engineering Manager @ Moderne