Poshyvanyk named Association for Computing Machinery Fellow

Denys Poshyvanyk, Chancellor Professor of Computer Science at William & Mary (Photo courtesy of Denys Poshyvanyk)

The following story originally appeared on the website for William & Mary's School of Computing, Data Science and Physics. - Ed.

Denys Poshyvanyk, chancellor professor of computer science at William & Mary, has been named an Association for Computing Machinery (ACM) Fellow, one of computing's highest honors, reserved for the top 1% of ACM members for outstanding technical achievement and service to the field.

Poshyvanyk is recognized for developing deep learning and software analytics methods that transformed software engineering research and practice. He was among the earliest to show that artificial intelligence (AI) could automate core software engineering tasks, from code generation and impact analysis to code review, documentation, clone detection and program repair.

His early work in these fields led to being awarded the "Most Influential Paper" award at both the ACM/Institute of Electrical and Electronics Engineers (IEEE) Mining Software Repositories and the ACM/IEEE Automated Software Engineering conferences in 2025.

Current research

Today, Poshyvanyk's work tackles the risks and realities of AI-generated code. One project aims to open the "black box" of large language models by defining a theory of causation that explains how LLMs generate their code. Better understanding of LLM-generated outputs would allow developers to use LLMs more responsibly and make more informed decisions, especially in high-stakes, security-critical systems.

Poshyvanyk notes that LLMs are trained on vast open-source data that can include security vulnerabilities. Without insight into how outputs are produced, those risks can propagate into real-world software. Additionally, particular prompts may result in outputs that contain security vulnerabilities.

"Our work helps developers better understand the how and why behind AI-generated code, leading to safer, more responsible outcomes," Poshyvanyk stated.

A second initiative, in collaboration with the William & Mary Law School, aims to help developers and organizations that utilize AI-assisted coding avoid legal compliance pitfalls.

"In an era where LLMs generate significant amounts of code, developers must understand the legal boundaries surrounding how that code can be used, adapted and shared," Poshyvanyk explains.

Because LLMs are trained on vast amounts of data, they can sometimes produce code that closely mirrors material governed by restrictive licenses. To address this, Poshyvanyk's team is developing an automated tool - essentially a live assistant embedded within the integrated development environment - that traces the origin of generated or imported code, identifies its licensing terms and guides developers on how to reuse it appropriately and legally in their own projects.

Impact on William & Mary

For Poshyvanyk, impact extends beyond scholarly achievement. After more than 17 years at William & Mary, he says the most rewarding part of his work is mentoring students.

"My lab is highly collaborative. I bring together students with complementary strengths, so they learn from one another," explained Poshyvanyk. "They push me to explore new ideas and take risks that often pay off."

Poshyvanyk is not only advancing the future of AI-driven software engineering, but he is also cultivating the next generation of researchers and innovators who will carry that work forward, ensuring that W&M remains at the forefront of responsible, high-impact computing.

Eva Kalajian

• share by email
• share on Facebook
• share on X
• share on LinkedIn

• share by email
• share on Facebook
• share on X
• share on LinkedIn

Related Stories