Production ML and analytics for retail banking operations.
Data scientist with 19 years in retail banking. I build production ML, SQL/Python pipelines, and Tableau dashboards that help teams monitor performance, prioritize work, and make better decisions. Android apps in my spare time.
Selected work
Public-data analytics and ML projects that show how I move from raw data to useful decisions.
CFPB Complaints Dashboard
Tableau dashboard on five years of CFPB complaint data, surfacing a structural gap in monetary relief between P2P and card disputes. Live on Tableau Public.
Loan Default Predictor
LightGBM model that estimates default probability on Lending Club personal loans, with a live Streamlit app you can poke at. AUC 0.71.
NBA Player Archetypes
K-means on per-100-possession stats finds eight player archetypes in the 2024-25 season. None of them line up with the traditional five positions.
Outside the day job
Two native Android apps shipped on my own time.
Stupid Small
An Android app that defeats procrastination by using AI to break overwhelming tasks into stupidly small micro-steps. Focus timer, streak tracking, and celebration animations included.
Niner
Native Android Sudoku app with five difficulties, five game modes, a daily streak puzzle, and a teaching hint engine.
Selected roles
The short version of the story behind the resume.
Data Scientist, AVP
- Built an XGBoost workflow that identifies operational errors 10x more often than random sampling, taken through validation, production deployment, and Tableau monitoring.
- Built the case-sampling tool QA reviewers use day to day; model scores drive risk-targeted review with a random control sample for ongoing validation.
- Build Tableau dashboards leadership uses to track operational trends, risk patterns, and team performance.
Sr. Business Analytics Consultant, AVP
- Owned analytics for relationship-operations strategy, supporting decisions for line-of-business leaders.
- Built SQL, PySpark, and Hive pipelines on Cloudera, orchestrated with Oozie, that fed downstream reporting and modeling work.
- Top 3, two years running, in PNC's enterprise data science competition: multi-class LightGBM (2024), Cox proportional hazards survival model (2025).
Business Analytics Consultant and earlier roles
- Built the first analytics function on multiple teams that previously had none.
- Replaced legacy Excel/VBA reporting with Python and SQL pipelines.
- Grew from reporting and collections roles into full analytics ownership.
Core stack
Tools and methods I use to move from raw data to business decisions.
Languages and platforms
Modeling and analytics
Visualization
Let's connect.
Always up for a conversation about data science, analytics engineering, business intelligence, or product work. Reach out anytime.