Pooya Rostami Mazrae

I am a software engineering researcher specializing in the intersection of machine learning, automation, and large-scale software ecosystems. I am currently a PostDoc in the Distrinet Group in KU Leuven Group T working on the secure use of agentic AI in software development. I have completed my PhD in Software Engineering at the University of Mons (UMONS), Belgium, where I investigated the evolution, quality, and reliability of GitHub Actions automation workflows. During my doctoral research, I have collaborated with Radboud University, and other research groups, producing peer-reviewed publications, datasets, and open-source tools.

Research: My current research focuses on secure use of agentic AI in software development in the context of Belgium companies in Flanders. As part of the COOCK project, this work examines the growing gap between the rapid adoption of agentic AI tools (single-agent and multi-agent, collaborative or adversarial) in industrial software teams and the corresponding maturity of their security practices. Through structured interviews with industry professionals, the study investigates key themes including prompt injection and supply chain risks, insecure code generation, CI/CD integration of AI tooling, and organizational governance and compliance around AI-assisted development. The goal is to characterize how Flemish software organizations are - or are not - adapting their security practices to keep pace with agentic AI adoption, and to produce empirically grounded guidance for more secure integration of these tools into real-world development pipelines. Moreover, My previous work focuses on analyzing and improving software development processes through data-driven methods. I study the evolution of automated DevOps pipelines, evaluate CI/CD ecosystems, and develop techniques for understanding and supporting large-scale workflow automation. I combine quantitative empirical analyses (mining GitHub data) with qualitative research (surveys, interviews, coding frameworks) to generate actionable insights for both researchers and practitioners. My works and collaborations has been published in leading venues including ICSME, MSR, EMSE, and Springer Nature, and has received distinctions such as the ICSME 2022 Distinguished Paper Award and the MSR 2022 Best Hackathon Paper Award.

Education: I hold a PhD in Software Engineering from the University of Mons (UMONS), Belgium, and an MSc in Software Engineering from Sharif University of Technology, where I developed machine-learning-based approaches for automating issue–commit link recovery. I earned my BSc in Computer Engineering from K. N. Toosi University of Technology, with a thesis on applying AI techniques to disease-gene detection. My academic trajectory integrates software analytics, machine learning, and empirical research methods.