Patrick Ding Email & Phone Number

Redmond, Washington, United States

• Evaluated double machine learning for estimating device specific causal effect of operating system update using AB test data• Processed terabytes of device telemetry data using internal SQL‑like tools and Spark to clean, merge, and engineer features for analysis• Shared problem findings throughout the team and to senior leadership, including a presentation for the Chief Product Officer

Lehi, Ut

• Implemented nested Monte Carlo estimators of KL divergence between MCMC, variational Bayes, and MAP posteriors of state space modelsusing R, Stan, and C++• Evaluated machine learning models for predicting loss of accuracy of approximate inference algorithms for state space models • Demonstrated viability of model for predicting when MAP gives accurate estimates

Greater New York City Area

• Guided investment decisions in Emerging Markets Multi Asset Portfolio with more than one billion dollars in assets by developing factor regression model for emerging market equity and foreign exchange returns• Guided investment decisions for variance swap sleeves with tens of millions of dollars in assets by building Matlab tools for back testing timing and selection strategies and developing factor regression model for predicting variance swap returns• Expanded firm’s risk management and volatility forecasting capability by modeling option implied volatility surface with factor regression model• Enabled firm to fulfill more stringent requests for proposal by creating realistic option simulator in Matlab supporting complicated trading, expiration, and hedging behavior with multiple assets• Discovered errors in competitor’s analysis of a trading strategy using option simulator , leading group to take on more successful call selling strategy• Impacted firm’s option rolling implementation by discovering excess returns dependent on expiration date using option simulator• Accelerated team’s analysis and model building for other funds , including Real Return Portfolio with more than one billion dollars in assets by developing Matlab and SQL tools for data loading and formatting• Communicated model results to team through daily updating Excel reports by interfacing Matlab COM automation and SQLserver databases

Greater New York City Area

Researched currency exchange rate volatility strategies to investigate income and protection potential• Collected and analyzed currency options data using Matlab and Excel• Examined returns from selling currency variance swaps calculated from options prices• Backtested factor-based timing strategies involving currency variance swaps

Doctor Of Philosophy - Phd, Statistics

Bse, Operations Research And Financial Engineering, 3.58

Activities and Societies: Daily Princetonian Design; Princeton University Sinfonia; Princeton University Orchestra; Princeton University.