Artificial Intelligence is rapidly transforming the way software is designed, developed, and maintained. At the AISE Lab at PUC-Rio, we investigate how AI can be leveraged to improve software engineering practices while addressing emerging challenges related to software quality, developer productivity, documentation, human factors, and trustworthy AI adoption. This talk presents an overview of our recent research on AI-driven software engineering, spanning the development of intelligent tools, curated datasets, and empirical studies that support both academic research and industrial applications. Our work includes automated documentation generation using Language Models (LMs), hybrid approaches for code smell detection that combine static analysis with machine learning and LMs, platforms for collaborative data annotation, large-scale mining of software repositories, and infrastructures for integrating heterogeneous software engineering data sources. Beyond technical advances, we also explore the human dimensions of software engineering, investigating developer emotions, well-being, diversity and inclusion, and the impact of AI-assisted development on software teams. These efforts have resulted in publicly available datasets, reproducible research infrastructures, and award-winning scientific contributions that foster transparency and collaboration within the software engineering community. The presentation concludes with lessons learned from building AI-powered software engineering ecosystems and discusses open research challenges toward creating more reliable, explainable, sustainable, and human-centric AI systems for the future of software development.
Website URL: http://www.inf.puc-rio.br/blog/professor/juliana-alves-pereira/
Juliana Alves Pereira is participating in the iVisit Program:
IRISA has funding available to invite foreign researchers (for short-term visits).
Annual call for applications