Andrew Choy

Predictive Pricing Analyst @Fiserv

US

Delivered $135M+ incremental revenue by architecting automated pricing optimization algorithms that transformed 3+ months of manual merchant analysis into sub-5-minute insights. Deployed production-grade Python infrastructure (NumPy, Pandas, SciPy, PySpark) on Palantir Foundry, integrating SQL/Spark data pipelines with Power BI executive dashboards—scaling seamlessly across SMB portfolios.Core Achievements:• Automated Pricing Pipeline: Architected Python ETL system using logarithmic, exponential, and polynomial (3rd/5th degree) curves with numerical integration (SciPy) to determine optimal pricing limits based on real-time merchant CTM. Natural log models provide asymptotic price-capping and embedded risk management—reducing pricing errors by 95% while maintaining ethical boundaries.• Operational Transformation: Accelerated pricing analysis from 3+ months to <5 minutes, enabling real-time adjustments across SMB merchants as portfolios and CTM values fluctuate. Leveraged PySpark on Foundry for distributed processing and Power BI for stakeholder visualization.• Retention Optimization: Reduced SMB churn through precision pricing that identifies the \"goldilocks zone\"—maximizing profitability without triggering attrition thresholds.Strategic Innovation:• Statistical Library Infrastructure: Building centralized Python repository (Gamma, Poisson, Log-Normal distributions) that eliminated 40+ hours redundant development per project. Enables parallelized analyses across merchant segments, transaction volumes, and retention cohorts using SQL/Spark queries.• Predictive Segmentation: Distribution-fitting algorithms identified 3 distinct SMB subpopulations, enabling targeted strategies and proactive retention interventions for at-risk segments.• Ethical Pricing Framework: Mathematical guardrails prevent predatory pricing while optimizing merchant lifetime value—proving algorithmic fairness and profitability are complementary.

Colorado College

Bachelor's degree