Antonio Mastropaolo
William & Mary, VA·Computer Science
AI systems that write and repair computer code now assist programmers daily. These large language models for code (LLMc) can generate code, fix bugs, write documentation, build tests and review changes.
- Fields
Computer Science
Engineering
- Topics
large language models
LLMs for code
LLMc
- Grants
- 1 known grant, through May 2027
What they work on
- Test whether filtering low-quality training examples with task-specific software engineering metrics lowers the computational cost of training large language models for code while preserving their learning ability.
- Apply model compression, particularly quantization (storing model weights at reduced numerical precision), to shrink code models and their resource consumption without degrading performance.
- Extend these efficiency techniques beyond code generation and summarization to a wider set of software engineering tasks such as bug fixing, documentation, testing and code review.
- Establish a structured methodology for assessing sustainability-focused optimizations and produce technical recommendations for energy-efficient AI software engineering tools.
More about the research
Training and running them takes enormous computing power, which means high energy use, high cost and substantial carbon emissions. This project asks how to make such tools cheaper and less resource-hungry without making them worse at their jobs.
Why it matters
Teams that rely on AI coding assistants could run them at lower cost and with less energy use and carbon output. Concrete guidance on data filtering and quantization could let developers deploy smaller models for bug fixing, testing and code review on modest hardware. Course material drawn from the work aims to prepare future software engineers to adopt these tools sustainably.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Lingming Zhang
University of Illinois Urbana-Champaign, IL·Computer Science
Train LLMs jointly to generate synthetic software issues and to resolve them, using a self-play loop in which the model learns to create the issues that most improve its own bug-fixing ability.
Juan Zhai
University of Massachusetts, MA·Computer Science·Early-career
Build a repository-level dataset linking natural-language documentation, formal behavioral specifications and code, with an automated pipeline for reproducible evaluation of specification accuracy.
Milos Gligoric
University of Texas at Austin, TX·Computer Science
Develop a language and framework for writing fine-grained tests that check individual statements and code fragments inside functions, and evaluate it on open-source projects.
The school
William & Mary: 11 labs on PI Pages, $95M a year on research, 67 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.