CV

BLOG

• 关于OWASP常见Web漏洞的调研报告(update: 2024-10-28)
• 关于OWASP-TOP10-LLM的调研报告(update: 2024-11-13)
• Claude 3.5与ChatGPT越狱提示(update: 2024-11-20)
• 大语言模型的对抗性攻击（Lilian Weng）(update: 2024-12-07)

CONTACT

• Tel: (+86)13516305590
• Email: jamiepw@163.com
• GitHub: github.com/JamiePW

EDUCATION

• 2023.09-2026.06: M.S. in Cyberspace Security, Beihang University
• 2024.01-2024.06: Exchange student, Université catholique de Louvain
• 2019.09-2023.06: B.E. in Information Security, Beihang University
  • Overall GPA: 3.85/4
  • Rank: 2/50

RESEARCH & PUBLICATION

• Wu Yuhan, Lü Jiqiang. Applicability of capsule networks in side-channel analysis[J]. Journal of Cybersecurity, 2026, 4(2): 1-14. DOI:10.20172/j.issn.2097-3136.260301.

• Dissertation for M.S.: Template Power Analysis Based on Neural Networks
  • Tutor: Prof. Lu Jiqiang
  • Abstract: To address the issues of blind model selection and poor generalizability in security evaluations of cryptographic implementations, this study investigates the adaptation mechanisms of neural networks for template-based power analysis. Through exhaustive search and bias-variance decomposition theory, it is demonstrated that Capsule Networks suffer from a mismatch between their spatial-semantic objectives and the requirements of temporal feature extraction, resulting in performance at only 75% of that achieved by CNNs while requiring five times as many parameters. Furthermore, a transfer attack scheme based on Transformer pre-trained foundation models is proposed. By employing flattened projection and autoregressive strategies to adapt the model for classification tasks, the proposed method achieves key recovery with a single power trace in cross-device scenarios (achieving 65% accuracy, reaching the theoretical optimum). Full fine-tuning yields a 372% performance improvement over tuning the classification head alone, validating the efficient transfer potential of general-purpose large models in side-channel analysis.
• Undergraduate Graduation Design(Thesis): Research on Implementation of High-Order Correlation Power Analysis of AES Implementation with Combined Countermeasures
  • Tutor: Prof. Lu Jiqiang
  • Abstract: In this paper, we studied an AES implementation that combines countermeasures such as affine masking and shuffling. We reproduce the power consumption model and perform computational optimization to implement higher-order correlation power analysis in 12 different scenarios, and evaluate the results at a finer granularity. We also propose a method to calculate the effectiveness of the attack results.
  • Code: github.com/JamiePW/ASCADv2-Extracted-Analysis

INTERNSHIP

• Internship at Sinovation Ventures (2025.05-2025.08, Beijing)
  • Responsible for daily IT support and issue resolution, promptly responding to and handling emergencies.
  • Independently designed and developed a backend monitoring system to meet the management needs of company servers and other network equipment, and handled its daily maintenance. code: github.com/JamiePW/mySimpleMonitor
• Internship at BITMAIN (2024.10-2025.01, Beijing)
  • Security Operations: Audited over 40 daily/monthly security hosting reports, authored and revised SOP documents totaling approximately 30,000 words for security operations, and maintained the software whitelist.
  • Security Testing: Participated in security testing for internal systems, with hands-on experience addressing OWASP TOP 10 web vulnerabilities (e.g., SQL injection, XSS, CSRF, SSRF).
  • Technical Research: Researched offensive and defensive techniques for large language model (LLM) security, producing 5 in-depth reports totaling over 40,000 words (see my BLOG).
• Internship at Institute for Network Sciences and Cyberspace of Tsinghua University (2022.06-2022.08, Online)
  • Tutor: Associate Prof. Zhang Han
  • Work during internship: I read and studied more than 20 papers in top journals or conferences related to my major, and gave presentations about typical cases. The total code amount of the project is about 2000 lines, and the effective code amount of a single file can reach up to 600 lines.
  • Code: github.com/JamiePW/tpds-socket
• Term Project of Machine Learning: Implementation of a Deep Learning Side-Channel Analysis Scheme for Boolean Masking Protected AES
  • Tutor: Associate Prof. Zhang Xiaoming
  • Abstract: Conduct side-channel analysis using MLP and CNN models based on TensorFlow and scikit-learn frameworks on the ASCAD open-source dataset, and compare attack results under different leakage models.
  • Work during internship: Experimental implementation and result analysis, as well as writing parts of the report.
  • Code: github.com/JamiePW/simple-ASCAD-DLSCA
• Term Project of Network Security Innovative Experiment: Security Vulnerability Exploitation and Defense for Small Networks
  • Tutor: Prof. Mao Jian
  • Responsibility: I built Linux VMs and network environments, implemented specific attacks and defenses in them(buffer overflow, weak password cracking, network scanning, SQL injection, etc.).
• Term Project of Digital Circuit and Systems: FPGA-based SM4 algorithm implementation and optimization
  • Tutor: Prof. Guan Zhenyu
  • Responsibility: Implementation(Verilog) and simulation(Modelsim) of SM4 cryptographic algorithm.

PROFESSIONAL SKILLS

• Languages
  • English: Fluent, IELTS Academic 7.5
  • Mandarin Chinese: Native
• Programming
  • familiar with C/C++/Python
  • Learning TensorFLow/PyTorch
  • beginning at SQL/Assembly Language/Golang/Rust
• Network Tools
  • beginning at OpenSSL/GunPG/OpenVPN/Nmap/Wireshark/netwox/hping3/HashCat/Snort

SOCIAL ACTIVITIES

• 2024.11-now: Member of Haidian Ballad Singers Association of Beijing
• 2020.04-2026.06: Member of Tianxiao Quyi Troupe of Beihang University

HONORS & AWARDS

HOBBIES

Xiangsheng(traditional Chinese crosstalk), pop music, sci-fi movies and novels