FIRESim
A privacy-first FIRE retirement simulator. Models portfolio growth, safe withdrawal rates, and retirement timelines. All computation is client-side — no data ever leaves the browser.
View projectEverything I've shipped
Apps, data tools, and experiments — built for fun, learning, or actual use.
A privacy-first FIRE retirement simulator. Models portfolio growth, safe withdrawal rates, and retirement timelines. All computation is client-side — no data ever leaves the browser.
View projectA browser-based platformer built with Phaser 3 and TypeScript. Fast-paced arcade mechanics, hand-drawn assets, and a chaotic difficulty curve that gets interesting fast.
Play itExploring economic mobility patterns across Arkansas counties using Opportunity Insights data. Compares upward mobility, income percentiles, and poverty rates against national benchmarks.
Read the analysisTime series forecasting for Ecuadorian grocery chain sales using LightGBM. Built for the Kaggle Store Sales competition — feature engineering on date, promotions, and store metadata.
View projectAmes housing price regression using XGBoost with extensive feature engineering. Handles 80 features, missing data imputation, and categorical encoding for the Kaggle Ames dataset.
View projectThe classic Kaggle classification problem — logistic regression with feature selection and cross-validation. First competition project; more about learning the workflow than winning.
View project