Automation · 2026
Job Search Engine
A private search engine for one job seeker.
Scans four applicant-tracking APIs across 55 companies every morning, scores every posting against a five-part model of what I actually want, watches my inbox for recruiter mail and files it against the right application, then sends one notification.
The scoring model is the interesting part. A raw keyword match tells you a job exists; it doesn't tell you whether to spend an hour on it. Each posting gets scored on how close the work is to what I've done, my realistic odds given the years and degree requirements, the posted pay against my target, whether the company actually funds the tooling I build with, and whether the location and culture signals are survivable.
It measures a company's AI investment from its own job board rather than from reviews. How many open roles are AI-titled, what share mention AI at all, whether they name specific tools. Reviews tell you how a place feels. The hiring pattern tells you what is funded.
It deliberately doesn't apply to anything. Mass auto-submission is exactly the pattern fraud detection looks for, and volume was already the variable that failed. Discovery, scoring and drafting are automated; the submit button stays human.
