Chrissy (Liangzhen) Du Email & Phone Number

San Jose, Ca, Us

New York, Ny, Us

• Market Movements Prediction Using World News - Utilized topic modeling to group world news into five categories and analyzed the relation between topics and market movements. Explored different classification algorithms such as Logistic Regression, Bayes, Neural Network, Random Forest and SVM. Improved out-­of-­sample accuracy by 10% in comparison to the baseline model, with dimensionality reduction, lagged data and feature normalization as well as mitigated class imbalance problems.• Airbnb Listing Analysis ­- Distinguished the features that were most important in determining Airbnb host success using machine learning on both listing, host and review features. Analyzed review data with Natural Language Processing (NLP) to extract sentiments. Displayed recommendation using D3 on how other hosts could maximize earnings and bookings.• Customer Loyalty Classification - Queried and analyzed > 20 million credit card transaction data to identify high­ performing customer, using tools AWS, PostgreSQL, Pandas, Scikit­learn and Google API.

Chiyoda-Ku, Tokyo, Jp

• Developed SQL queries to pull data and run reports in Excel for risk assessments and capital forecasting as in­-house support of the Capital Management team• Developed scenario tests for various situations and conditions, in order to ensure that the company maintained necessary capital reserves• Lead process improvement projects of automation the data pull and report updating process through SQL and ETL. Reduce data delays and improve data accuracy by optimizing the monitoring and validation processes• Designed and developed dashboards/visualization in Tableau, presented the findings to senior management. Liaised and communicated expectations between the regulators and internal stakeholders in the bank

Cleveland , Ohio, Us

• Developed loss forecasting models, commercial credit risk rating models and consumer scorecard models by extracting data from MongoDB, MySQL and flat files using Python and R to support business decisions and identify risks and opportunities. Performed hyper-parameter tuning (Adam & SGD) to optimize performance • Analyzed loan performance and borrower behavior data, to figure out how to optimize credit decisioning strategy, collaborated with engineers to implement the solution• Develop digital marketing model using deep learning method (RNN/LSTM) with Keras and PyTorch. Conducted A/B testing, the model increases conversion rate by 25%• Performed weekly analysis and reporting on key trends to allow internal business teams and senior management to understand issues and identify actions to solve these issues

Master'S Degree, Statistics

Bachelor'S Degree, Financial Engineering