Xinye Zhao (Jasper)
I am a Ph.D. student in Computer Science and Engineering at the University of Notre Dame, advised by Prof. Joanna C.S. Santos. My current work focuses on AI4SE (AI for Software Engineering) and software engineering security.
My recent research centers on AI for software engineering security, especially agentic and reasoning-based methods that help developers analyze, validate, and improve the security of software systems. I am particularly interested in how multi-agent architectures and LLM-based reasoning can support security-critical engineering workflows such as secure code analysis, vulnerability assessment, and assurance for real-world domains including medical device software. My previous work spans systems for machine learning, including LLM KV cache optimization; AI for Science, including computational biology research on single-cell learning systems; and IoT network traffic analysis.
Before joining Notre Dame, I completed an M.S. in Electrical and Computer Engineering at Georgia Tech. Contact: or LinkedIn .
Research Directions
AI4SE and Agentic Development. I investigate how LLM-based agents can help developers understand code, navigate complex engineering tasks, and make software-development workflows more reliable.
Software Engineering Security and Program Analysis. I build methods that combine static analysis, semantic reasoning, and human-centered validation to improve the precision and practical usefulness of security analysis.
Systems for ML and LLMs. I am interested in the systems foundations that make intelligent models efficient and dependable in practice, from scalable inference to reusable representations.
Collaboration. I am always interested in connecting around AI for software engineering, software security, program analysis, and reliable LLM systems. If our research overlaps, please reach out through the contact options above.
Recent Highlights
News“STRATA: A Security Tactic-Aware Framework for LLM-Based Vulnerability Repair” is now available as an SSRN preprint.
News“Towards Budget-Aware Early Candidate Selection for Vulnerability Analysis” accepted to EXPRESS 2026.
News“TriVul: Improving Precision of Static Vulnerability Detection through Multi-Agent Reasoning” accepted to the Static Analysis Symposium (SAS 2026).
Published in IJCNLP 2025: SemShareKV: Efficient KVCache Sharing for Semantically Similar Prompts via Fuzzy Token Matching.
Joined the University of Notre Dame as a Graduate Research Assistant.
Graduated from Georgia Tech with an M.S. in Electrical and Computer Engineering.