Modern AI agents are stepping beyond chatbots and into the developer’s toolbox—capable of planning, reasoning, and fixing real bugs in real codebases. In this talk, we’ll dive into Google’s latest research on agent-based program repair, where large language models (LLMs) are used to autonomously diagnose and patch complex bugs at scale.

We’ll explore Passerine, an experimental repair agent built to operate inside Google’s own development ecosystem, and how it performed on a new dataset of 178 real-world bugs. You’ll learn:

🛠️ How AI agents go from bug reports to working code

 📊 Why enterprise bugs are different from open-source ones — and what that means for industry adoption of bug repair

 🚀 What these results mean for the future of coding and developer productivity

Whether you're curious about LLMs, developer tools, or just want to see AI take a crack at real software engineering problems at the scale of Google, this talk is for you.

📍 Location: TU Wien - FAV 01 (https://maps.tuwien.ac.at/?q=HEEG02#map)

📅 Date: Thursday, 15. Mai 2025

🕖 Time: 18:00–21:00

🍕 Perks? Free Pizza, Free Innocent Juice, Free Snacks & drinks, great people, top-tier AI content!

💸 Cost: Free! Limited capacity – please register early to secure your spot!

🎙 Meet our speaker:

Jürgen Cito is an Associate Professor at TU Wien (Vienna, Austria). He received his PhD from the University of Zurich and conducted postdoctoral research at MIT CSAIL. He has also been continuously engaging with industry as a visiting scientist and software engineer at Meta (in the Probability group) and was recently a visiting faculty researcher at Google (in the DevAI group). His current research spans AI for code, with particular focus on explainability of code models, and leveraging LLM agents for security testing and program repair.