Statistics & Data Science @ UCLA (Dec 2027) · Data Engineering + Data Science · Python · R · SQL
I work at both ends of the data stack — building the pipelines that move and validate data, and the models that make sense of it. Most of my work is about making data trustworthy before anyone builds on top of it: baseline audits, validation gates, and honest evaluation of whether a model is actually right.
- Data quality & pipelines — read-only audits, synthetic-data testing with planted-corruption answer keys, pre-deploy validation gates
- Entity resolution — confidence-scored matching to replace fragile free-text joins
- Predictive modeling & evaluation — regression, win-probability, and calibration (Brier scores)
Python (pandas, scikit-learn, NumPy, PyTorch) · R (dplyr, tidyr, ggplot2, Shiny) · SQL · Git · GitHub Actions
- ucla-football-data-quality — Data-integrity audit, synthetic-data testing, and a 50-check pre-deploy validation gate for a production sports-science pipeline (sanitized).
- nba-win-probability — Live win probability from play-by-play: game-state features + MLP (PyTorch), Random Forest & XGBoost, evaluated by Brier score.
- wbb-transfer-scouting — Multi-season scraping (On3 + Sports-Reference), EDA, and a filterable React/TypeScript scouting dashboard.
- ncaa-basketball-spending — Do bigger athletic budgets buy success? Multi-season linear modeling in R; published on Bruin Sports Analytics.
- wnba-betting-market-analysis — Are sportsbook odds well-calibrated? Calibration curves and Brier scores.
- ufc-weight-jump-analysis — Weight-class jumps vs. fighter ROI; published with WagerWire.