Huiyu Chen

Shopee - Marketplace Intelligence - Chatbot

• Leveraging LLMs to enhance Shop Chatbot, which serves millions of stores across 8 key markets. Improved customer satisfaction (CSAT 70%), boosted order conversion (+2.00%), and increased deflection rate (+10.00%), reducing live agent dependency. Specialized in fine-tuning, deployment, and optimization, developing Topic Recognition, Product QA, and Sales Guide modules to automate inquiries and optimize recommendations.

Tencent News - Recommendation Algorithm Center

• As a Recommendation Algorithm Engineer Intern, I worked on optimizing recommendation models and improving feature selection to enhance content personalization and user engagement.

Developed a Prompt-Guided Multi-Classification Model to route customer inquiries efficiently, achieving 80.44% in-topic accuracy and increasing deflection rate by 10.72%, reducing unnecessary agent transfers.

Designed an LLM-powered Product Q&A to answer buyers' product-related inquiries, ensuring buyer satisfaction exceeded 70%. Handled 33.9% of chatbot traffic, improving CSAT and reducing live agent escalations.

Implemented Function Call-based query parsing for personalized product recommendations, boosting conversion rate by +2% and increasing recommendation CTR by 5%, enhancing the shopping experience.

Optimized the double tower model structure, filtered and added effective features, resulting in a 0.8% increase in click-through rate (CTR).

• First-Class Scholarship ; GPA: 3.71/4.0
• Research & Publications: "Sleep Stage Classification Using Single-Channel EEG based on CNN and Transformer"
• Granted Patent: "Human-Computer Interaction Method and System Based on Electrooculography (EOG) Signals"

• Outstanding Graduate Award ; First-Class Scholarship ; GPA: 3.75/4.0 (Top 5%)
• Research & Publications: ""Near-infrared spectrum of coal origin identification based on SVM algorithm"
• Competitions & Awards: Honorable Mention, MCM/ICM Contest in Modeling (COMAP)