Live product
Baseball Analytics Hub
→A PuckCast-style baseball section: recent games, win probability movement, leverage moments, model receipts, and a path toward daily team analytics.
MLB Stats APISupabase-backed game stateWE / WPA / LI
CS @ Illinois · Backend · Data Systems · Applied ML
Software systems for sports, markets, and messy real-world data.
I build projects that move past demos: ETL pipelines, normalized databases, API layers, model inference surfaces, validation loops, and clean product interfaces.
Current focus
Baseball decision simulator
API layer
Stable endpoints over normalized MLB game state
ETL
Scheduled refresh jobs with sync logging
Model surface
Win expectancy, leverage, recommendations, metadata
Product
Manager replay and PuckCast-style analytics direction
Baseball
A serious analytics product, not just a chart.
Inspired by the way PuckCast turns NHL model outputs into a daily product, this section frames baseball work around repeatable data pipelines, accountable projections, and useful decision surfaces.
Finance
Research code with risk controls and validation.
Market projects that show Python analysis, portfolio construction, event studies, and backtesting discipline.
Research notebooks
AQR Hold the Dip Replication
→Python replication of AQR-style market timing work using SPY data, drawdown-triggered rules, costs, trend-following comparisons, and event studies.
PandasEvent studiesStrategy diagnostics
Trading competition
UChicago Portfolio Optimization
→Daily portfolio construction system using covariance shrinkage, risk-parity baselines, alpha overlays, turnover controls, and holdout validation.
Risk modelsBacktestingTurnover-aware weights
Systems
Other applied software work.
Additional projects across automation, robotics, collaborative tools, and product engineering.