Summary
Haoyang Liu is a research-focused software engineer with a Ph.D. in Computer Science and Engineering from HKUST and four years of experience specializing in security, formal methods, and LLM-assisted vulnerability analysis. Currently a Research Assistant at HKUST, he developed a systematic approach that integrates large language models with traditional data race detectors to reduce false positives in Linux kernel concurrent vulnerability detection and probed LLMs’ limits with tailored prompts. His work spans automated evaluation frameworks for LLMs in CTF/CVE scenarios (built during a UW–Madison internship) to formal-methods-based smart contract analysis using Spin and Mythril. Comfortable bridging theory and tooling, he has a knack for customizing analysis modules and building unified automated environments that reveal nonobvious failure modes in real-world code. Based in Wuhan, China, he combines deep academic training with practical engineering to push the boundary of automated security discovery.
4 years of coding experience
B.S., Cyber Science and Engineering, B.S., Cyber Science and Engineering at Huazhong University of Science and Technology