Matthew Fay Email & Phone Number

New York, Ny, Us

Conversion propensity modeling for ad targeting:• Increased lead conversion rate by 30% in Google Ads campaigns spanning $3M annual marketing spend.• Trained gradient boosting model to predict lead conversion with >85% AUC using CRM, engagement, and product usage data.• Built automated ETL, feature engineering, and inference pipelines using Python, Snowflake, and DBT to generate propensity scores for new leads and upload to Google.Revenue forecasting with timeseries ML:• Developed revenue forecasting model used by go-to-market leadership for strategic decision-making, account prioritization, and sales quota setting.• Forecasted revenue across 3000+ enterprise customers with median errors of 12% at an annual horizon and 5% at a quarterly horizon.• Designed recursive graph traversal method to handle revenue attribution and improve training data continuity for customers who switched account IDs.

New York, Ny, Us

New York, Ny, Us

• Forecasted time series of US rent price by zip code using XGBoost, ARIMA, and Lasso models, with mean error under $100/month across 1300+ zip codes. Integrated data sources including Google Trends, bike sharing services, and the US Census. Reported to real estate intelligence firm Markerr with assessment of the predictive value of each data source and best candidates for inclusion in their internal models.• Predicted house sale prices in Ames, Iowa with 95%+ accuracy and developed an interactive house-flipping dashboard allowing virtual modification of a home and visualization of projected changes in price. Integrated geospatial data from OpenStreetMap and deployed models including Lasso regression, gradient boosting, support vector machines, and stacked ensembling.• Scraped OpenTable data on all restaurants in NYC to extract 25 features per restaurant, visualized trends in cuisine preferences and user ratings, and identified key factors in maximizing OpenTable bookings using linear regression.• As assistant instructor, led sessions to reinforce concepts spanning Python and R coding, advanced statistics, gradient descent, and machine learning algorithms. Mentored student data science projects in direction and technical implementation.

New York, Ny, Us

• Designed and delivered series of lectures on university-level chemistry topics to premedical students.• Tailored communication based on learning styles, knowledge levels, and personalities of audience.

Chapel Hill, Nc, Us

• Modeled drug metabolism using nonlinear regression and drew conclusions on experimental outcomes via statistical A/B testing.• Coordinated with multiple specialized laboratories to bridge knowledge gaps as needed for different stages of the project.• Drove project planning and direction, and communicated research findings as primary author of publication.• First ever undergraduate presenter accepted to Gordon Research Conference on Drug Metabolism.

Bridge Program, Computer Science

Md-Phd Candidate

Bachelor Of Science (Bs), Chemistry