Penn Recruit

What it is

penn-recruit builds a database of high school wrestlers from tournament result pages, then uses TypeSafe's Jev model to classify pages, tier prospects and find the right school counselor. It drafts transcript requests on the day a prospect's NCAA contact window opens. One command (penn-recruit run) executes the whole crawl, classify, ingest, tier and outreach flow, and everything is idempotent and cached.

The problem

A recruiting staff cannot manually track thousands of prospects across tournament sites or know when each one becomes contactable.

How it works

1. A polite crawler (robots.txt honored, 2 seconds per host) and a result-line parser turn pages into matches, wrestlers and graduation years.
2. Jev classifies pages (high school, youth, college), scores prospects on a five-level scale, finds the school site and counselor emails, and picks the caseload counselor for a given last name.
3. An eligibility module computes each athlete's June 15 contact window; drafts are re-checked right before sending and held for review when school matching is uncertain.
4. SQLite database, a Typer CLI, an HTML dashboard, and systemd service and timer files for scheduled runs.

Stack

Disciplines

For your business

A prospect-discovery and outreach pipeline that finds leads on the open web, scores them with AI and drafts contact-window-aware messages for approval.