Hi —
I write production data systems from a cattle farm in rural Alabama. By day, that's senior data engineering at Just Insure — pricing platforms, Bayesian and ML pipelines, GCP infrastructure, and the LLM agents I'm increasingly building on top.
Twelve years ago I started in insurance audit — SQL scratchpads and quarterly reviews. Over the years I rebuilt that function into a software platform, spent a stretch on the West Coast in Oregon, picked up an applied-math M.S. from Johns Hopkins, and eventually came home to Alabama — by then on the engineering side, building the systems rather than auditing them. Off-hours: 20+ head of cattle, a garden, and side projects I treat as methods to expand the craft — a public dashboard for the Clay County livestock auction, another for Alabama economic indicators, and a Python library powering an MLB analytics platform my friends and I use through the season.
End-to-end modern-data-stack analytics platform for Major League Baseball statistics. Dagster orchestrates scheduled ingestion from Statcast, Baseball Reference, and FanGraphs into DuckDB; dbt models apply dimensional modeling; Superset surfaces self-serve dashboards. Asset lineage runs from raw API pulls through marts and visualization.
Live fantasy-league analytics product built around a RAG architecture over a custom multi-source pipeline. Ingests Sleeper rosters, NFL news, injury feeds, and weekly context stats with per-source cadence and rate-limit handling. LLM prompts are grounded against assembled context to produce per-team weekly recaps and trade analyses.
Public-facing economic-monitoring dashboard tracking state and federal macroeconomic indicators via the FRED API. Built on the Observable framework, deployed as a static site on GitHub Pages with automated refresh so non-technical stakeholders can track Alabama's trajectory against national benchmarks.
Python client library unifying baseball data from Statcast, Baseball Reference, and FanGraphs behind a single interface. Handles rate limiting, schema normalization across heterogeneous public APIs, and caching — used as the ingestion layer for the MLB Analytics Platform.
MCP server exposing BigQuery schema metadata to LLM-powered tools (Claude Code, Cursor). Lets AI coding agents reason over warehouse structure when generating SQL or dbt models, reducing hallucinated column names and types and accelerating analytics-engineering work.
Built the pricing platform and led a domain-oriented data mesh migration on GCP. Owned end-to-end systems including a real-time rate-comparison engine, an AI sales assistant, and MCP servers feeding warehouse metadata to LLM coding tools.
Designed and built the Continuous Audit Program — a data warehouse and Python anomaly-detection platform that replaced manual reviews with continuous evaluation. Grew the internal Data Analysis Framework into a multi-team analytics platform.
Architected the enterprise data warehouse and led the org-wide Tableau rollout. Built the Rewards Platform for the member loyalty program.
Built the Audit and Data Analysis Framework — an internal application that automated fraud-detection queries across policy, claims, and billing systems.