AI Austria Meetup

AI Austria is back in Innsbruck for the 2nd edition of our local community meetup series!

This meetup continues our growing community series in Tyrol — focused on knowledge exchange, networking, and collaboration around all things AI. Expect deep technical insights, inspiring talks, and an open space for everyone passionate about artificial intelligence — no matter your background.

Agenda
🕡 6:15 pm Open Doors, Arrival pm
🕡 6:30 pm // Start of Event, Welcome
🕡 6:45 pm // 🎙 How anomalous is your code? – Johannes Troppacher, Dynatrace

Modern software systems grow fast, become deeply interconnected, and often evolve beyond what any single engineer can fully grasp. Traditional static analysis helps, but it is usually rule-based and focused on syntax. What if we treated code as data — and let machines learn its structure and oddities?

Beyond hype and buzzwords, we explore how unsupervised graph machine learning can detect anomalies in real-world codebases. By modelling code as a graph, enriching it with graph algorithms, and turning structure into embeddings, we can cluster similar components and surface those that do not quite fit. Not bugs. Neither good nor bad. Just unusual — and often a hint at design smells like the God Class.

You will see how to reproduce these results and try them on your own code using my code-graph-analysis-pipeline. Along the way, I will share challenges I faced as a software engineer and how I approached solving them.

Whether you are curious about graph databases, machine learning beyond buzzwords, or software design anomalies, this talk will give you new ways to look at your code and ask: How anormal is it, really?

🕡 7:30 pm // 🎙 Where Are All the Smart Robots? A Quest For Sample Efficiency in Reinforcement Learning – Samuele Tosatto, LFU

Reinforcement Learning holds an immense potential to solve real-world problems. We witnessed agents learning to beat grandmasters at chess and Go, and world champions in video games. Despite that, we are not yet surrounded by smart robots doing our house chores and helping us in our everyday lives. Why is that?

While most of today's AI (chatbots, video generators, ...) are trained on vast amounts of data, this cannot be easily done for real-world agents. I argue that one way to shorten the gap between reinforcement learning and real-world applications is to make it more efficient. This talk will discuss how we can make reinforcement learning more efficient by enhancing its statistical engine and by refining its working assumptions to better fit the real world.

🕡 8:15 pm // Community Events - What's happening in Innsbruck? A short overview of interesting events happening by the community for the community.
Then Food, Drinks & Networking until
🕡 9:15 pm // end

## AI4Industry Meetup – "Measurable Business Impact" Edition

Artificial Intelligence cannot longer be a hype topic — without real business value & impact.

As part of the AI4Industry series, this meetup focuses on exactly that challenge. The central question of the evening is how leaders can introduce, govern, and scale AI in a way that creates real, measurable business value - beyond buzzwords and isolated pilot projects.

## Program

From 5:00 PM – Check-In

### 5:30 PM – Keynote Insights from Industry, Academia & Consulting