Felix Yuzhou Sun

Risk Management: Customer Default Risk Prediction

• Lead the team of four using machine learning to predict the default rate of an account.
• Extracted and merged data from the database with SQL.
• Performed feature screening on 202 features with Major Dimension-Reduction techniques (PCA), resulting in 74 numeric and 79 categorical features.
• Cleaned and imputed the data of 25k credit card transaction records. Prepared the processed data for machine-learning logistic regression by oversampling.
• Trained and tested data with Random Forest (scikit-learn). Further improved the predicting power by Gradient Boosting (XGBoost). Evaluated the result based on the Confusion Matrix.
• Made the group presentation and visualized our methodology.

Impact Analysis on Head Start Program, Philadelphia

• Identified confounding variables and researched for causal relations between explanatory variables and intervention outcomes with path-analysis using the Multi-Mediator model.
• Implemented Repeated Measurement ANOVA and ANCOVA between the treated and controlled groups with R.
• Visualized the analysis of program effectiveness with publication-quality statistical listings and plots with R(ggplot).

Financial Data Analyst Intern (Remote)

• Engaged in the team's due diligence for the equity investment of Chunmi Technology, a small household appliance enterprise in the Xiaomi ecological chain; analyzed data and performed valuation forecasting.
• Collected market data with web scraping (Requests). Merged and cleaned the data and performed comparability analysis in Chinese Household Appliances Industry between target companies and companies in the reference group with Python.
• Established linear regression model forecasting the valuation growth in the next five years based on the current market share and market growth in the past five years.
• Wrote the Analytical Report, visualized the data, and performed group presentation.

Real-time Quant Trading System (In progress)

• Integrated APIs for seamless real-time market data acquisition.
• Created trading indicators based on volatility, momentum, and moving averages, leveraging Pandas and NumPy.
• Built a QT-based interactive GUI for real-time monitoring of trading metrics and algorithm performance.
• Deployed custom algorithms with machine learning to predict stock price movements and validate trading signals.
• Implemented custom metrics and simulations to rigorously assess trading strategy effectiveness and adapt to market conditions.

University of Pennsylvania Graduate School of Engineering

Philadelphia, PA

Computer and Information Technology | GPA: 3.9

University of Pennsylvania Graduate School of Education

Philadelphia, PA

Statistics Measurement & Research | GPA: 4.0

Heidelberg University

Heidelberg, BW, Germany

German Studies & History

Passed Level I Exam of the CFA Program

Passed Level I Exam of the FRM Program

PCAP – Certified Associate in Python Programming