Alan Du

Introduction

I am currently pursuing a degree in Financial Engineering at Columbia University, with a focus on carving out a unique path for myself on Wall Street. As an aspiring quantitative researcher, my core passion lies in translating complex mathematical concepts into practical and deployable code. I enjoy exploring friction and ineffiency of the market using a variety of tools, including Machine Learning, Statistical Modeling, and Quantitative Finance, in order to derive meaningful insights.

I have prior experience working as a quantitative researcher at HKU CFID (Center of Financial Innovation and Development, the University of Hong Kong), where I successfully navigated a range of challenges related to statistical arbitrage. Throughout my two years there, I was involved in every aspect of strategy development, from ideation to live execution.

My interests include trekking, jogging and poker.

Work

• Constructed one delta-hedge arbitrage strategy on commodities based on the observation of a significant intraday volatility change pattern between U.S. and EU/Asia trading hours. Employed grid search to fine-tune timing and select appropriate option pairs. Incorporated dynamic execution and stop-earn algorithm to increase volume. Earned 2.2% average monthly return with 7.8 Sharpe.
• Created one rule-based speculation strategy of on-shore and off-shore (Singapore, HK) China equity index futures. Adopted co-integration to explore asynchronous correlation patterns. Conducted quantile analysis and isotonic maximum rank test to determine close out timing conditional on opening price change. Realized average 4.1% monthly return with 5.2 Sharpe.
• Produced a Spot-Fut arbitrage strategy by investigating mean reversion between XAU and COMEX gold. Revealed a non-monotonic decreasing pattern of the price gap between spot and future gold. Adopted a self-designed predictor to infer the intraday price gap peak. Backtested the predictor and achieved annualized 9% return with 4.7 Sharpe ratio conditional on a < 5% risk-free rate.
• Connected Python strategies with C-based China Comprehensive Transaction Platform to ensure < 0.05 second execution delay. Transferred ticker-level binary data flow from 4 major Chinese futures exchanges by integrating C shared libraries into Python classes.

• Developed a systematic trading model demo on non-parametric arbitrage timing. Used meta-ML techniques e.g., joint maximization and iterative testing to empower automatic rule and model selection for different assets.
• Tested the micro price predictability of innovative methods e.g., isotonic maximum rank correlation, kernel and spline regression. Achieved 70%-win rate on average for the next 10 seconds on certain commodities.

• Identified 10+ revised profit factors of convertible bonds based on underlying stock price, volume and bond terms, e.g., computed the quantile of current convertible premium rate on historical data to find factors have IC over 2%.

Research

• Purposed non-parametric kernel estimator for average treatment effect test under a semi-supervised learning framework.
• Applied the estimator to help the sector selection algorithm for A shares and improved annualized return by 2% in back-tests.

• Analyzed the hypothesis statistical power of all 3 existing multivariate statistical testing methods for high-dimensional datasets.
• Proved the approximated equality in three methods by simulation with over 100 price and volume features.

Education

Relevant Coursework: Trading Algorithm, Monte-Carlo Simulation Methods, Statistical Analysis and Time Series, Financial Engineering Continuous Time Models, Optimization Models and Methods for Financial Engineering, Machine Learning for Financial Engineering, Stochastic Models, Foundations for Financial Engineering

Relevant Coursework: Probability Distributions and Bayesian Probability, Causal Inference, Stochastic Process, High Dimensional Multivariate Data Analysis, Machine Learning, Econometrics, Financial Derivatives, Linear Algebra, Programming in C/C+, R Programming, Python