Case Study · Job Search Automation

AI Job Engine

The wide-net counterpart to Autopilot Job Hunt: 12 portal scrapers feeding one pipeline that scores fit with Gemini Flash, tailors the resume per job description, and applies with one Playwright click.

12 portal scrapers1000+ jobs tracked100% automated
⎋ View on GitHubDiscuss this project

01Why it exists

Job listings are scattered across incompatible portals — LinkedIn, Greenhouse, Lever, Indeed, Naukri, Wellfound, Instahyre, ArbeitNow, Hirist, Remotive, WeWorkRemotely, Hacker News. Each has its own markup, session quirks, and noise. Searching them by hand means missing most of the market.

02How it works

  1. Scrape 12 portalsDedicated scrapers per portal normalise listings into one SQLite store — dedup across sources included.
  2. Score with Gemini FlashEvery job is fit-scored against the profile: fast, cheap, and consistent triage across 1000+ tracked roles.
  3. Tailor the resumeHigh scorers get a dynamically tailored resume per job description before any application goes out.
  4. One-click applyPlaywright drives Easy Apply flows end-to-end; a Streamlit dashboard tracks status of every application.

03Results

12portal scrapers
1000+jobs tracked
100%automated
10×faster hunt

04Stack

PythonGemini FlashPlaywrightBeautifulSoupStreamlitSQLite