← Projects
CFB Sharp Private
A college football analytics platform with an LLM data pipeline and an MCP server for AI analysts.
What I built
- Built 17 scheduled Python jobs that feed an eight-factor model.
- Wrote an LLM pipeline that turns injury news into structured data, validated with Pydantic.
- Shipped an MCP server that lets AI analysts query data and run experiments, with every write requiring human approval.
- Saved every AI analyst session as a transcript, which keeps sessions available for later evaluation.
How it works
At a glance
| Type | Independent |
|---|---|
| Status | Private |
| Stack | Python, FastAPI, TypeScript, Pydantic, MCP, Linux |
Launch
='[cfb_sharp.xls]Board'!A1This cell links to another workbook, and that workbook is private.
Open linked workbookLast updated