SQL, Python, R, Statistics, Time Series, Machine Learning

Yiyun Wu

Washington, DC **********@*******.*** 973-***-**** LinkedIn

Talented and self-motivated data analyst with significant experience in data manipulation and validation, statistical analysis and modeling. Skilled in analytics tools, such as SQL, Python, R, Toad, SAS, MS Excel. Master’s Degree in Quantitative Finance from Rutgers University SKILLS

Programming: Python, SQL, MATLAB, R, C/C++, VBA, Shell, SAS Tools: Toad, Jupyter Notebook, Putty, JIRA, SPSS, EVIEWS, MS Excel, MySQL, SharePoint Models: Data Manipulation/Validation, Generalized Linear Models, Time Series, Statistics, Machine Learning, Monte-Carlo Simulation PROFESSIONAL EXPERIENCE

Fannie Mae Washington, DC

Data Analyst Oct 2019-Present

Core member participated in the project from design to implementation of the whole process for the successful database migration from RDW to EDW and redesign of Credit Enhancement (CE) profile. Contributed to business requirements gathering, data transformation analysis, code development, data testing and production release reconciliation

• Redesigned a whole new CE profile generating process by researching on database structure and streamlining complicated and redundant data derivation logics. My proposed design was adopted in helping achieve a more robust data product

• Collected and reported important development requirements from cross-functional teams. Designed user stories based on data rule analysis with clear vision for code development and refactor

• Worked out complex data derivation logic to implement ETL jobs by using Shell and SQL scripts for CE profile development

• Performed system testing and data reconciliation to troubleshoot data issues. Created and executed test cases to ensure both functional and non- functional requirements were implemented correctly with desired data quality for product release plans going smoothly

• Became a subject matter expert in the team for CE profile and developed presentations and solution documentation from business and technical perspective, including creating workflow diagrams for data derivation logic and CE profile generating process QuantConnect Seattle, Washington

Quantitative Developer Intern Sep 2019-Oct 2019

• Researched on trading strategies and API of QuantConnect developing platform to support clients for troubleshooting their trading strategies

• Designed and coded trading strategy algorithm of Technology ETF by using NLP method on QuantConnect Python-based developing platform Yansheng Technology Shanghai, China

Quantitative Researcher Intern Jun 2018-Aug 2018

• Researched, implemented and modified futures CTA strategies via Python-based third-party trading platform

• Coded and implemented Reversion Strategy and Pair Trading Strategy on different future contracts using technical indicators with Python

• Performed back testing with historical data, calculated risk ratios and did mock trading on VN-PY trading platform

• Applied daily K-line data and ATR indicator to filter for concussion interval, improved performance in back testing and efficiently managed risk EDUCATION

Rutgers University Newark, New Jersey

Master of Quantitative Finance Aug 2017-May 2019

Courses: Object-Oriented Programming, Data Mining, Time Series, Econometrics, Optimization, Stochastic Calculus, Numerical Analysis Competition: CME Trading Challenge 2018 Top 10%; Teaching Assistant of Data Mining course (Feb-May 2019) Hunan University Changsha, China

Bachelor of Financial Management (Financial Engineering), Mathematics Minor Sep 2013-Jun 2017 Courses: Linear Algebra, Probability, Econometrics, Stochastic Process, Financial Engineering, Financial Risk Management Leadership: Class President (2013-14); Office Minister of Astronomy Club in Hunan University (2014-15) RESEARCH PROJECTS

Risk Management: Default Behavior Analysis Mar 2019-May 2019

• Preprocessed 850k+ loan records with pandas of Python. Applied linear regression, pivot table to analyze factors and filtered for useful variables

• Developed Logistic Regression, Random Forest, Naïve Bayes and K-Nearest Neighbors models to find out the relationship between default status and its factors and predicted default behavior in an efficient way

• Used PCA method to do dimensionality reduction, evaluated and compared models with confusion matrix and ROC curve. Found Random Forest was the most suitable method due to unbalanced dataset

• Applied SMOTE to expand minority class samples, which significantly improved minority class recognition by Random Forest R Programming: Bitcoin Price Prediction with ARIMA model May 2018

• Applied both ADF (Augmented Dickey-Fuller) and KPSS (Kwiatkowski-Phillips-Schmidt-Shin) tests to determine the stationarity of Bitcoin’s close price time series from 04/01/2015 to 05/28/2017

• Did stationarity transformation with second difference and obtained AR (p) and MA (q) values using ACF and PACF methods

• ARIMA model was used in a dynamical way to predict the close price of Bitcoin in the daily basis

Contact this candidate