AI agents are rapidly evolving from simple assistants into autonomous systems capable of reasoning, maintaining context, and interacting with complex environments. In this meetup, we’ll explore two practical applications of modern agentic AI in software engineering.
The first talk dives into building stateful AI agents that work with real-time data, covering the architectural patterns and challenges behind creating responsive, context-aware systems. The second talk showcases how agentic AI can be applied to understanding and analysing complex legacy applications — helping uncover program flows, dependencies, and hidden system knowledge that would otherwise take countless hours to explore manually.
Whether you’re building AI-powered applications, modernising enterprise systems, or simply curious about the future of intelligent software, this meetup offers hands-on insights into how AI agents are transforming the way we interact with technology
THIS MEETUP WILL BE HYBRID. | Zoom Link will be available shortly before the meetup
-
6.30 PM | Doors open. Grab some snacks & drinks.
-
7.00 PM | Welcome & Intro – Start of hybrid meetup
-
7.15 PM | Building Stateful AI Agents with Real-Time Data, Gregor Bauer (VP Customer Engineering at CrateDB)
-
~8:15 PM | Hosty - The next level with Agentic AI, Martin Binder (Lead IT Consultant at QAware)
-
afterwards: More snacks & drinks, and a lot of time for networking with all attendees and speakers.
Building Stateful AI Agents with Real-Time Data, Gregor Bauer (VP Customer Engineering at CrateDB)
Most AI demos are stateless. Real systems are not.
Modern AI applications increasingly depend on continuously changing operational data: events, metrics, logs, transactions, sensor streams, and user activity. This creates a new architectural challenge: how do we build AI agents that can reason over live data while remaining scalable and observable ?
This talk explores the architecture of stateful AI systems and agentic applications. We will cover:
-
streaming-first AI architectures
-
real-time context enrichment
-
combining vector search with structured queries
-
memory layers for AI agents
-
latency and scalability considerations
-
operational lessons from large-scale data platforms
The session provides practical guidance for engineers designing production-ready AI platforms.
Hosty - The next level with Agentic AI, Martin Binder (Lead IT Consultant at QAware)
To migrate legacy systems, you need to understand these IT dinosaurs, at least partially. So code analysis up to a certain level is unavoidable. What used to take months, sometimes years without LLMs is now doable in a few days with Gemini and friends. And yet: you still have to drive the analysis yourself, stitch the LLM's partial results into a coherent picture and fight the context window. With old source files thousands of lines long, that part really hurt.
Agentic AIs like Claude Code are not just another AI feature. They change how you approach unfamiliar systems — fast, and without the headaches that LLM chat workflows used to come with. Teaser: an analysis that took us a few days under "yay!" cheers is now done in minutes.