002
Job Match Agent
A local AI-assisted job search workspace that finds listings, scores fit, tracks applications, and publishes a sanitized dashboard.
01 — Why I built this
Job searching manually is exhausting: the same listings appear across boards, most do not match my actual skills, and the work of tracking applications quickly turns into scattered links and notes. I built a private local workspace that reads my resumes, searches multiple boards, scores listings for skill alignment and seniority fit, tracks applications, imports outside URLs, and publishes a sanitized public dashboard without exposing private notes or contacts.
02 — How it works
Resume routing, multi-source job fetch, AI scoring, SQLite persistence, local workspace, email, and sanitized public dashboard
The Search Run tab is the command center. Typing run starts the full browser-launched search pipeline, while live cards track scraped listings, matches, stretch roles, fit roles, and filtered results against the active target city map.
03 — What was hard
- Dual-resume routingI am applying to both Data Analyst and Software Engineering roles, which use different resumes. The agent routes jobs to the right resume by title so DA and SWE matches are scored against the relevant profile.
- Private local data vs. public proofThe local app stores contacts, notes, follow-ups, resume versions, and application statuses in SQLite. The generated public dashboard intentionally strips those details while still showing aggregate progress and top matches.
- Turning discovery into a workflowFinding jobs was only half the problem. The app now supports saved/applied states, follow-up dates, external URL import, run history views, editable target cities, and salary estimates so the search becomes a managed pipeline.
04 — By the numbers
05 — Links