Runmeng Zhai

Enterprise Data Analytics Pricing Analytics

• Supported new credit card products launch under limited historical data by leveraging legacy portfolio benchmarking and overlay-based modeling to estimate risk and profitability.
• Built monthly automated early read reporting (~300K new accounts) for senior leadership, tracking key portfolio metrics (e.g., %IBAR) to provide forward-looking insights for model calibration and strategy alignment.
• Developed automated Financial back-test framework using financial metrics (e.g., NIBT) to evaluate model performance and refine credit and pricing strategies.
• Leveraged AWS SageMaker (job-based pipelines) + Python to build scalable data processing and visualization workflows, reducing manual reporting effort by 30%+.
• Presented bi-weekly ad hoc analyses and model insights to cross-functional stakeholders and leadership, translating complex modeling outputs into actionable business recommendations.
• Led end-to-end development of pricing and profitability models for education refinance loan products, including model design, validation documentation, and implementation, improving predictive accuracy by 12% and reducing runtime by 41.6%
• Built elasticity-based pricing optimizer tool incorporating macro factors (SOFR, Fed cycle) and borrower segmentation (FICO, degree, loan term) to support pricing strategy decisions and volume forecasting.

Credit Risk Analytics

• Developed and enhanced Probability of Default (PD) models, benchmarking Logistic Regression vs. XGBoost, improving performance from 68% → 77%, and deployed scoring logic into production systems.
• Designed and implemented Expected Loss (EL) pipeline using Python and Google Cloud data sources, improving data consistency and supporting pricing and credit strategy decisions.
• Built risk monitoring dashboards (Tableau / Python) to evaluate credit strategies across borrower segments, improving efficiency of credit policy optimization.

• Built a KNN Recommender System to identify comparable stores for agents to get better estimate of the sales projection and reference the market demographic to find out the optimized market strategy and got store approval rate increase to 40% from 30%.
• Developed sales forecasting models (Logistics Model) with different business strategy, wrote Python codes to implement the models into enterprise system.
• Analyzed market trends and customer data to optimize sales strategies, resulting in a 15% increase in quarterly revenue.