I build data platforms and AI agents with SQL and Python. At Good360, I'm a Senior Data Scientist working across the warehouse, semantic layer and tools people use to get answers from the data.

• I own and operate a Synapse warehouse and its ingestion pipelines across 10 source systems and hundreds of millions of rows.
• I built a governed semantic layer, four MCP servers and five AI agents for analytics, project management, disaster monitoring and Power BI documentation.
• I integrated a donation channel that accounted for 60% of the organization's volume. The resulting cleanup added $1.42B in recipient attribution and raised Form 990 Schedule I coverage from 53% to 94%.
• I took the analytics platform from zero to 1,000+ monthly views across 20 departments and secured $1.5M in grants for the team's technology development.

More on that work at twhite.dev.

• PRTL turns home displays into mobile-controlled art and information frames.
• Under Par Fantasy Golf runs salary-cap PGA leagues with live tournament scoring.
• DocPBI documents Power BI reports, tracks model changes and writes definitions back into the model. The SaaS also includes an API-powered assistant and an MCP connector.

Scout Skills is a local skill ranker and curated library for coding agents, including Claude Code, Codex and Cursor. It searches skill descriptions on disk so agents can find the instructions they need without loading the whole library.

Orator turns recordings or transcripts into scored speech reports with highlighted delivery and word-choice feedback. It combines verbatim speech recognition with timing and acoustic analysis, and can suggest stronger wording where someone hesitates.

Boondock is an offline browser map for dispersed camping on US public land. It combines geospatial data with the rules set by each land manager to answer where someone can camp.

dbt/DuckDB baseball warehouse models pitch data through staging, intermediate and mart layers, with an incremental fact table, model contracts and data tests in CI. Includes sample data for a local build without credentials.

baseball-analytics brings together NCAA scouting tools used by University of San Diego coaches, a dbt and DuckDB warehouse, and MLB pitch and game models. It includes pitcher and hitter apps, run-expectancy forecasting, batter-pitcher matchups and Monte Carlo game simulation. The scouting apps and transfer portal board have been relied on by USD's back-to-back WCC regular season championship teams.

Fantasy Football combines college-to-NFL prospect modeling with dynasty roster analysis and Monte Carlo trade simulation. Its prospect research compares position-specific college-production and draft-pick models using time-ordered backtests and draft-class bootstrapping.

• Kicker-Analysis: weekly NFL kicker forecasts with season-forward backtests and baseline comparisons.
• CBB: March Madness models testing whether team statistics add information beyond tournament seed.
• WanBattedBall: a CatBoost batted-ball notebook from Nick Wan's Kaggle competition.
• Bart: compares model lines with Kalshi prices for college basketball and NHL games.

For project details or to get in touch, visit twhite.dev.