Biography

George K. Thiruvathukal is a professor of computer science and the department chairperson at Loyola University Chicago and is a visiting computer scientist at Argonne National Laboratory. His research interests include high-performance computing, distributed systems, software engineering, machine learning, and computer vision. Thiruvathukal received a Ph.D. from Illinois Institute of Technology.

SRP Project Title

Software Engineering for Science in the AI Era

HPSF Project

Topical Areas

Applied Computer Science; Artificial Intelligence and Intelligent Systems; Computer Science; Open Source Software; Sociology; Software Engineering; Statistics and Probability

Abstract

This project investigates how researchers across scientific disciplines are reusing modern AI technologies—including pre-trained models (PTMs) and intelligent agents—in their research workflows. Although deep learning and related AI methods have transformed computational practice, training large models or building advanced agents remains costly and technically complex. As a result, many research communities increasingly rely on existing models and agent frameworks, yet we know relatively little about how this reuse is occurring outside computer science. Our study combines manual analysis with large-scale automated methods to develop a clearer understanding of AI reuse in practice. We are examining open-access scientific articles, datasets, and software projects to determine when AI technologies are reused, which types are most prevalent, and how researchers adapt or repurpose them for domain-specific tasks. Processing and analysis will be performed on advanced high-performance computing resources at Argonne National Laboratory to support large-scale document and data processing. We also classify reuse strategies using established software engineering taxonomies and analyze the relationship between reuse and scientific impact. Students joining the project will participate in empirical analysis, automated document processing, and user-study components to understand how scientists adopt, adapt, and evaluate emerging AI tools in their work.

Desired Skills

We welcome students with solid Linux and Python skills. Above all else, we value problem solving skills, effective communication and collaboration skills. We will consider students from any major, especially those in science and engineering programs that emphasize learning at least a little bit of computer science!

Lightning Talk Title

AI usage across scientific disciplines using large-scale analysis

Keywords

software engineering; mining software repositories; mixed-methods; statistical methods; Python; Linux; open source; open access; high-performance computing; distributed systems