Resume

Carnegie Mellon University

Master of Science in Intelligent Information Systems (MIIS)

• Language Technologies Institute, School of Computer Science

The Chinese University of Hong Kong

Bachelor of Engineering in Artificial Intelligence: Systems and Technologies

• GPA: 3.82/4.00, Major GPA: 3.92/4.00, Rank: 1st of 39
• First Class Honours, ELITE Stream, Dean's List (2020 - 2024)
• Graduate-level Courses: Machine Learning Theory, Foundations of Optimization, Pattern Recognition
• Specialized Stream: Large-scale Artificial Intelligence - Theory and Systems

QuadSentinel: Sequent Safety for Machine-Checkable Control in Multi-agent Systems

Yiliu Yang*, Yilei Jiang*, Qunzhong Wang, Yingshui Tan, Xiaoyong Zhu, Sherman S. M. Chow, Bo Zheng, and Xiangyu Yue

LogicAsker: Evaluating and Improving the Logical Reasoning Ability of Large Language Models

Yuxuan Wan*, Wenxuan Wang*, Yiliu Yang, Youliang Yuan, Jen-tse Huang, Pinjia He, Wenxiang Jiao, and Michael R. Lyu

AI Engineer

RedotPay

• Co-developed and productionized an AI customer-service agent serving approximately 3,500 daily conversations with knowledge-grounded responses, MCP tools, SOP workflows, conversation analytics, and human handoff.
• Compared with the legacy Dify bot, the agent increased CSAT from 71.6% to 84.0%, AI-only resolution rate from 79.2% to 85.6%, and deflection rate from 70.6% to 81.9%, saving an estimated 50 support hours per day.
• Designed and built a 909-case automated regression framework for tool-using workflows, validating required and prohibited MCP calls with mocked backend states across KYC and SOP scenarios.

Research Assistant

Multimedia Laboratory, The Chinese University of Hong Kong

• Developed QuadSentinel, a multi-agent guard system translating natural language policies into machine-checkable rules for robust runtime safety enforcement.
• Achieved 91.5% accuracy on the AgentHarm safety benchmark across prompt injection, jailbreak, and tool misuse attacks, outperforming prior state-of-the-art methods.
• Supervised by Prof. Xiangyu Yue.

Final Year Project

Department of Computer Science and Engineering, CUHK

• Explored the generalization ability of GANs by evaluating the effectiveness of Inception-based metrics such as FID and KID.
• Demonstrated that Inception features may fail to capture overfitting, highlighting the need for alternative evaluation methods without neural network reliance.
• Final Year Project supervised by Prof. Farzan Farnia.

Summer Research Intern

Purdue University

• Conducted research on machine unlearning by leveraging knowledge distillation and continual learning to mitigate membership inference attacks.
• Developed efficient unlearning frameworks that preserve model utility while enhancing data privacy.
• Supervised by Prof. Ninghui Li.

Research Assistant

Department of Computer Science and Engineering, CUHK

• Developed the core code of the LogicAsker framework, an automated system for evaluating and improving logical reasoning in LLMs using atomic logic rules.
• Implemented fine-tuning procedures that significantly improved reasoning accuracy in LLMs.
• Supervised by Prof. Michael R. Lyu. Published at EMNLP 2024.

Summer Research Intern

Department of Computer Science and Engineering, CUHK

• Developed continual learning techniques for dynamic Knowledge Graph embeddings using Elastic Weight Consolidation and edge replay sampling.
• Achieved efficient incremental updates with minimal forgetting, adaptable to any embedding method, earning the Best Project Award.
• Supervised by Prof. James Cheng.
• [GitHub] [Poster]

Summer Research Intern

Department of Computer Science and Engineering, CUHK

• Developed APM, a scalable C++ system for fast approximate graph pattern matching using neighborhood sampling.
• Designed algorithms to support general patterns with auto-optimized sampling strategies, outperforming existing methods.
• Supervised by Prof. James Cheng.
• [GitHub] [Poster]

International Collegiate Programming Contest (ICPC)

CUHK

• ICPCID
• 2022-23 Hong Kong Regional (Bronze)
• 2022-23 Shenyang Regional (Bronze)
• 2021-22 Macau Regional (Bronze)
• 2021-22 Kunming Regional (Bronze)
• Coach: Prof. Siu On Chan

Olympiad in Informatics (OI)

• OIerDb
• 2019 CCF CSP-S (First Prize)
• 2019 IOI CTSC (Bronze)
• 2019 CCF Winter Camp (Bronze)
• 2018 NOIP (First Prize)

• ELITE Stream Scholarship
• Dean's List 2023-24
• Talent Development Scholarship
• APEC Scholarship
• Gold Award for Outstanding Academic Performance
• Scholarship for Academic Excellence

• HKSAR Government Scholarship
• ELITE Stream Scholarship
• Dean's List 2022-23
• Scholarship for Academic Excellence
• Bronze Award in ICPC Regional Hong Kong Site
• Professor Charles K. Kao Research Exchange Scholarship

• Bronze Award in ICPC Regional Shenyang Site
• ELITE Stream Scholarship
• Dean's List 2021-22
• Scholarship for Academic Excellence
• Bronze Award in ICPC Regional Kunming Site
• Bronze Award in ICPC Regional Macau Site

• ELITE Stream Scholarship
• Dean's List 2020-21
• Talent Development Scholarship
• APEC Scholarship

• First Prize in CCF CSP 2019
• Second Prize in CPhO in Provinces 2019
• Second Prize in CMO in Provinces 2019
• Bronze Medal in IOI CTSC 2019
• Bronze Medal in CCF Winter Camp 2019
• First Prize in NOIP 2018