Hong Kong, Hong Kong SAR
A data scientist and quantitative researcher with expertise in the end-to-end development of machine learning-driven trading strategies. Proficient in deriving predictive signals and building statistical models in Python, C++, and R. Proven track record in data processing, alpha research, model tuning, and implementation for both cross-sectional and time-series approach.
- Researched 50+ alpha factors for long-short equity strategies, deploying regression models (WLS, Lasso, LightGBM) to extract pure alphas (orthogonal to risk factor models) and conduct signal selection/fusion. - Built several scalable data pipelines to process TB-scale tick data, via SLURM clusters using Bash Script & Linux Tmux Commands. - Developed a Python data engineering library using lightweight algorithms (e.g., JIT, Parallel, Vectorising) for efficient factor mining. - Concluded and deployed 2 new actionable alphas, with cross-checked correlations and high-performance metrics (bps, segregation, ICIR, turnover). - Weekly submitted documentation on work progress and findings on the latest industrial or academic papers.
- Engineered and maintained data pipelines using Python requests and remote webhooks to collect and structure real-time crypto trade data. - Fetched and cleaned auxiliary data, including news, official announcements, and social media, by web-scraping, LLM search, and NLP engineering. - Researched, backtested, and paper-traded an independent high-frequency momentum strategy, leveraging Ensemble Models (RF, LGB). - Assisted the development of a mid-frequency statistical arbitrage strategy in pair trading, specifically in Cointegration and Kalman Filters studies. - Weekly group evaluation meeting on work progress to drive continuous enhancement and report directly to CTO.
Supported research by collecting and cleaning up primary data. Maintaining organised project files and documentation for validation. Co-author of a PhD-level article presenting longitudinal study examining the relationship between ESG practices and financial performance at China Construction Bank.
Student Thesis Project for BSc Theoretical Physics at University of Glasgow Conducted in-depth analysis of b-quark jet substructure in ttbar dileptonic decays using C++ Monte Carlo simulations (CERN Rivet & Pythia system, Docker). Measuring key observable (N-Subjettiness, Les Houches Angularity, Energy Correlation Functions, C2 and D2 correlations) and comparing MC predictions with real ATLAS laboratory data. Established a solid selection criteria for the "Tag and Probe" in identifying valid b-jets.