Xinhai Zou

Thesis: Security Architecture for Industrial Control Systems

Thesis: Wireless transmitter fingerprinting; 14.56 (Cum laude)

Change to Master of Artificial Intelligence in 2021; 13.9 (Cum laude)

Thesis: In-situ observation of phase transformation during heat treatment of high strength stainless steel produced by Selective Laser Melting

• Partnered with a materials science lab to integrate AI techniques into domain research workflows.
• Developed an LLM-based assistant that wraps existing ML models into a conversational interface.
• Built a scalable pipeline for automated collection, parsing, and ingestion of large-scale academic literature from web sources.
• Designed and implemented data infrastructure for structured storage, retrieval, and downstream analysis of research data.

• Built and deployed a WeChat Mini Program game “Meme” that leverages real-world news to simulate public relations decision-making and scoring players' strategy.
• Developed core algorithmic components for “Raven”, an economics modeling and quantification product, achieving 100% return over a 3-month evaluation period (Jan–Mar 2026).
• Led end-to-end development of the Android application for “Raven”.
• Designed and documented internal workflows via SKILL.md for the OpenClaw framework, powering automation of financial reporting, work summaries, and planning processes for company.

• Supervised Master’s theses and participated in examination processes.
• Conducted research on Software Bill of Materials (SBOM), analyzing systemic limitations in current generation pipelines and proposing mitigation strategies.
• Developed machine learning-based methods for device fingerprinting across multiple domains, including IIoT devices and mobile phones.
• Investigated security risks in LLM-based multi-agent systems, focusing on both system-level and protocol-level vulnerabilities.
• Applied machine learning techniques to Physical Unclonable Functions (PUFs), studying modeling attacks and corresponding defenses for hardware security.