Hi, I'm Lavine.
I have a background in Mathematics and Applied Mathematics and am currently pursuing an MSc in the Department of Computer Science at City University of Hong Kong. My work focuses on two connected tracks: AI products and quantitative systems.
I like turning ambitious ideas into working systems from zero to demo, connecting technical judgment with product context, growth thinking, and stakeholder communication.
Recently, I've been exploring AI agents, data-driven products, and interactive worlds that make complex concepts easier to use, explain, and share.
Selected Work / Updated 2026
AI products, quantitative research, and interactive systems — each tied to a result, a working prototype, or a public artifact.
A relationship-centered AI product for preserving authentic life records and entrusting them to loved ones with consent and restraint — not digital resurrection.
2nd Place in Track · Hong Kong Physical AI Hackathon
An AI coding collaboration product that brings agent execution, shared project context, and team review into one live workflow.
First Prize · Agent Builder Hackathon, Shenzhen · Hosted by StepFun
A Python factor-research workflow for data cleaning, IC / IR signal evaluation, and competition return validation.
National Runner-up · Top 1%
An AI networking product that structures professional identity and helps users discover higher-value connections.
30-hour Hackathon Prototype
A game-based product that turns candlesticks, market sentiment, and trading strategies into an explorable learning experience.
AI-assisted Finance Learning Prototype
Current Builds / Quant systems in public
My current direction is to make quantitative research easier to inspect and reproduce: point-in-time inputs, explicit failure states, cost-aware validation, and public implementation notes.
Point-in-time screening
A fail-closed A-share moat screener with auditable evidence, explicit data boundaries, and reproducible JSON / Parquet outputs.
Cost-aware factor research
Five-session A-share reversal research with point-in-time controls, cost-aware backtesting, and reproducible validation.
Experience
Roles and proof points behind the projects above.
2026
Forbes China ESG Impact Summit
Supported executive reception, cross-cultural communication, and stakeholder connections across enterprise leaders and institutions.
2025 — 2026
PandaAI Quant Factor Competition
Built and validated quantitative factors with Python, achieving national runner-up and top 1% competition performance.
2025
Shenzhen XDF
Taught DSE mathematics, helping students build structured problem-solving methods and exam-oriented quantitative thinking.
2024
UPDF
Developed international content matrices and search strategies for overseas growth.
2021 — Present
Douyin
Built a data-informed content loop, with a single video reaching 40K+ views and 21K+ likes.
Education
2025 — Present
Department of Computer Science · City University of Hong Kong
2020 — 2024
Haide College · Ocean University of China
Capabilities
Two connected tracks: human-centered AI products and reproducible quantitative systems.
LLM Applications / Agents / Human-centered AI / Rapid Prototyping / Demo Design
Factor Research / Point-in-Time Data / Cost-aware Backtesting / IC / IR / Reproducible Validation
Python / TypeScript / JavaScript / React / Statistical Modeling
Git / GitHub / Codex / Claude / Evidence-led Iteration