Engineering in the Loop
Details
Java & Friends is back after a longer-than-planned break! Because most JVM engineers are currently busy keeping pace with AI, we’re (again) leaning into the "& Friends" part of our name. We’ve brought together three builders, starting with a former Java developer turned CTO, to share practical software engineering lessons on orchestrating AI agents, LLM workflows, deciding where humans belong in the loop, and how developer responsibilities and required skills are shifting.
THIS MEETUP WILL BE AN IN-PERSON EVENT 🕺
We're starting at 17:40 🕒
What to expect:
🗣️ Two talks, three speakers
👥 Meet old and new friends from the community
🍕 Free food
🎉 Free drinks afterwards
✔️ Just RSVP!
Agenda:
- 17:30 - Doors open
- 17:40 - Welcome by & and SQUER
- 17:50 - First talk
- 18:35 - Break
- 18:50 - Second talk
- 19:35 - Food, Drinks & Networking
- 21:30 - Doors closing
👂 Talks:
🗣️ Michael Bauer-Wapp
From Full-Stack Developer to Customer-Facing Orchestrator
🗣️ Victor Webersberger & Nuris Samardžić
Human in the Loop, but Which Loop?
🧑💻 Speakers:
Michael Bauer-Wapp is CTO at BauTask, a construction management platform built on Ruby on Rails, and an independent software consultant based in Austria. He has more than 10 years of professional software experience, starting with Java and Android and moving through test automation and full-stack web development. He holds an MSc in Software Engineering & Internet Computing from TU Wien, where he also taught software testing and quality for several years.
Talk description:
From Full-Stack Developer to Customer-Facing Orchestrator
I started as a Java developer and worked my way through full-stack roles. Today, as CTO of a construction software company, I spend much of my time orchestrating agents and working out what customers actually need.
That sounds like leaving engineering behind. My experience is the opposite. I can turn a customer problem into a precise requirement quickly because I know the software inside out, and agents can only build what someone has defined clearly.
In this talk, I'll follow real customer needs from the first conversation to shipped software. We'll look at which skills became more valuable, which mattered less, and what this means for developers early in their careers.
------------
Victor Webersberger has been a Software Architect, a Java Consultant, a Data Platform Architect and a Forward Deployed Engineer. Which one is current? Honestly, he's not sure, and it doesn't really matter: next week he'll be something else, and the week after that probably something that hasn't been invented yet.
What stays the same: He loves hard problems, messy systems and solutions that actually make it to production. Sometimes that means designing the architecture, sometimes writing the Java, sometimes building data platforms, and sometimes sitting next to the client until it works.
Most of Nuris Samardžić's work lives behind the screens people use every day and he doesn't really care where in the stack the problem hides. Over 18 shipped projects in AI, fintech, public services, and urban mobility, he's learned that production doesn't care what your job title is; it only cares whether the system holds up when real users hit it.
Today at andamp, he builds AI agents, the cloud infrastructure they run on, and the apps people use to talk to them. In practice, that means doing whatever gets an idea from a blank page to a tested release: co-architecting zero-credential agent platforms, red-teaming them himself or shipping a 45,000-line AI marketplace solo. He likes strict security, clean architectures, and code that survives contact with the real world.
Talk description:
Human in the Loop, but Which Loop?
Everyone says to keep a human in the loop. Almost nobody says which one. Our agentic workflow at andamp has multiple loops: planning, building, a review loop where reviewer agents send findings back to the implementation agents until the code passes and more. Put a person in every loop and you’re coding by hand, only slower. Take them all out and nobody notices when the work drifts. We’ll show where we keep people, where we let the agents run, and the questions we ask to decide. The model you put on each step matters too, because it changes how much a human has to check.
------------
Can't find what you like? Got something to say? 📢
Send us your thoughts in the comments.
Whether you’re a seasoned speaker or stepping onto the stage for the first time, we’d love to have you share your insights at one of our upcoming events. Interested in giving a talk?
Let us know!
Our formats:
• Full talks (~30 to 45 min.)
• Short talks (~15 min.)