Jürgen Cito

AI + SE Seminar Series (June 11, 2025. 11 am – 12 pm EST).

Evaluating Agent-based Program Repair at Google

Agent-based program repair promises end-to-end bug fixing by combining planning, tool use, and code generation via large language models. While prior work has focused on open-source benchmarks like SWE-Bench, we explore the viability of such approaches in an enterprise context using a curated dataset of 178 real-world bugs from Google’s issue tracker—78 human-reported and 100 machine-reported. We present Passerine, an agent adapted to Google’s development environment, and evaluate its performance. Passerine achieves plausible fixes for 73% of machine-reported and 25.6% of human-reported bugs, with 43% and 17.9% of these being semantically equivalent to the ground truth. Our results establish a baseline for agentic repair in industrial settings, highlighting key differences from public benchmarks.

Bio

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.

Visit: https://ipa-lab.github.io

EECS Upcoming Events: https://lassonde.yorku.ca/eecs/eecs-upcoming-events/

Zoom Registration: Here