Our beautiful Berlin QA community gathered again in early October, and even with autumn in full swing, around 70 people showed up for another great evening of ideas, demos, and honest conversations about quality.
I’m so damn happy to see how this community has grown: familiar faces who’ve been with us since the very first meetup, and new folks jumping right in. It really feels like we’ve built something special together: a real community, not just an event series.
I’m also incredibly proud that most of our speakers this time started out as regular meetup attendees, people who once sat in the audience and are now up on stage sharing their own stories, and present them extremely well!
Digital Twins 101
Slides are here.
We kicked off with Mitesh Patel taking us into the fascinating world of digital twins — virtual replicas of real-world systems used to test and improve everything from software-defined vehicles to AI-powered surgical robots.
Mitesh broke down what digital twins are, how they work, and why they’re changing the way we test and develop complex tech. He connected model-based engineering, simulation, and real-time feedback into a single picture — showing how this approach makes development faster, safer, and smarter.
I was extremely happy to bring Mitesh on stage: he’s a true scientist, and I honestly believe we in engineering need to rebuild a stronger connection with science itself. That’s how engineering was meant to be, grounded in curiosity, experimentation, and rigorous thinking.
A super insightful talk for anyone curious about the next frontier in testing.
Automated testing of a Large Language Model
This talk was special, being the first talk in our meetup’s history coming straight out of our Beyond Quality community research!
Anupam tackled one of the trickiest modern challenges: how do you automate testing for a Large Language Model like ChatGPT, when its responses aren’t deterministic? Most testing frameworks rely on predictability, but LLMs don’t.
His live demo showed how you can use one LLM to judge another’s output, basically, an AI acting as the QA reviewer. That sparked a ton of discussion on how to validate systems that are probabilistic by nature.
He also walked us through Retrieval-Augmented Generation (RAG), the same mechanism that lets LLMs answer questions based on uploaded documents without retraining.
This talk wasn’t just cool, it showed what our community is capable of when we collaborate, explore, and experiment together. And that community collaborative energy keeps going strong: I’m currently preparing two of my conference talks based on our ongoing community research: QA Principles in Hiring and Economics of Testing.
Slaying the Complexity Hydra: QA in Compose Multiplatform
Vlad took us behind the scenes at JetBrains, where he was the first QA engineer on the Compose Multiplatform team, starting from scratch with no processes, no quality definition, and no automation.
His story was about scaling both complexity and people. As Compose expanded from desktop to iOS and web, Vlad had to balance process design, automation, and team growth to keep quality sustainable.
Key lessons included the power of risk-based testing, knowing when to adjust tooling, and recognizing that QA isn’t about chasing perfection, but rather about keeping confidence high even as scope and technology evolve.
See you at the next meetup in Berlin on 2nd December!
Please follow our Quality Engineering Meetup Berlin group to hear about upcoming meetups in Berlin. We’ll also be experimenting with new cities around Europe — follow us on LinkedIn to be the first to hear about new locations.
