JULIA (YING) CHEN

• Developed and implemented diverse machine learning models (logistic regression, random forest, neural networks, XGBoost, and time series models) that improved the accuracy of forecasting recruiting feasibility, employee promotions, transfers, and attrition trends by 35%, contributing to more informed workforce planning
• Designed and maintained scalable data pipelines and predictive models, enabling real-time analytics of key workforce metrics, resulting in a 85% reduction in reporting time and enhancing data accessibility for stakeholders
• Created and deployed NLP models for topic modeling and sentiment analysis, improving the assessment of employee feedback and generating actionable insights that boosted employee engagement scores by 60% and strengthened organizational culture
• Collaborated with cross-functional teams (Finance, HR, Recruiting, and Operations) to devise effective headcount strategies, identifying key performance metrics and translating complex analyses into actionable insights
• Developed interactive dashboards and visualizations to present forecasting outcomes and metric analyses to leadership, facilitating strategic planning and prioritization that led to accelerated decision-making processes
• Engaged in ongoing research and applied advanced analytical techniques, driving continuous improvement of models and processes within the People Analytics team, resulting in a 20% increase in overall analytical efficiency

• Built deposit attrition random forest model to predict customer-initiated account closure (100MM+ records) and improved 3-month retention by 10%
• Developed credit decisioning origination system for SME products (credit card, term loan, line of credit, and reserve line) using 3000+ features. The system has successfully decreased 24-month default loss by 17%
• Mitigated charge-off loss in $2MM+ for line of credit and reserve lines by establishing a suite of behavior scorecards to monitor account status
• Applied machine learning techniques in consumer credit card fraud detection. The system’s accuracy was increased by 30% and annual fraud loss was decreased by 21%
• Launched A/B Test on checking account promotion landing page, which led to 23% increase in total number of new accounts opened online
• Established final model specifications (applied to all the models mentioned above) for the data engineering team, monitored/confirmed all data flows to the production platform, and conducted tests on the deployed models and attributes
• Proactively presented highly interactive data visualization results and actionable insights to stakeholders and ensured project progresses were in line with business goals

• Built cluster-based logistic regression models in SAS for credit card account origination and improved annual revenue by 6%
• Led score cutoff analysis for checking/saving accounts by comparing different origination channels and strategies. Created rule sets on top of the score, which brought an increase of $937K in annual revenue