Gold Rush

Status: Archived. This project was built to automate a personal job search; archived after the author accepted an offer. Code is preserved as a reference implementation.

An end-to-end personal job-search pipeline: scrape, dedupe, score, and surface fresh listings through a password-gated dashboard.

What it does

Runs twice a day on a local Mac via launchd, pulling fresh job postings from three sources:

1. Gmail IMAP — parses LinkedIn / Indeed job-alert emails
2. JSearch API (RapidAPI) — active search across aggregated boards
3. Hacker News "Who's Hiring" — scraped via a companion script

Each job is scored 0–100 against configurable preferences (title, location, salary, company), stored in SQLite, deduplicated across sources, and delivered via:

• A morning + evening digest email
• Immediate alerts for high-score ("platinum") jobs
• A web dashboard updated on every run

Architecture

┌─────────────────┐    ┌──────────────┐    ┌──────────────────┐
│ Gmail IMAP      │───▶│              │    │                  │
├─────────────────┤    │  gold_rush   │───▶│  SQLite (local)  │
│ JSearch API     │───▶│   .py        │    │                  │
├─────────────────┤    │  (scoring +  │    └────────┬─────────┘
│ HN Who's Hiring │───▶│  dedup)      │             │
└─────────────────┘    │              │             ▼
                       └──────┬───────┘    ┌──────────────────┐
                              │            │ jobs.json        │
                              ▼            │ meta.json        │
                       ┌──────────────┐    │ (exported)       │
                       │  Email       │    └────────┬─────────┘
                       │  digest      │             │
                       └──────────────┘             ▼
                                         ┌──────────────────────┐
                                         │  Vite + React        │
                                         │  dashboard (Vercel)  │
                                         └──────────────────────┘

Repo layout

.
├── gold_rush.py                 Main pipeline: sources → score → store → export → email
├── tpm_job_hunter.py            HN "Who's Hiring" scraper (runs earlier in the day)
├── run_daily_job_hunter.sh      Wrapper invoked by launchd
├── com.yourname.*.plist         launchd job definitions (twice-daily schedule)
├── config.env.template          Copy to config.env; fill in credentials + preferences
├── dashboard/                   Vite + React + TypeScript dashboard
└── docs (GOLD-RUSH-*.html/md)   Setup guides, patch notes, PRD

Setup

1. cp config.env.template config.env and fill in:
   • Gmail address + App Password
   • RapidAPI key for JSearch
   • Target titles, locations, salary floor, preferred companies
2. pip install -r requirements.txt (or install deps listed at top of gold_rush.py)
3. First run: python gold_rush.py
4. Schedule via launchd: update the .plist files with your paths and launchctl load them
5. Dashboard: cd dashboard && npm install && npm run dev

Full walkthrough in GOLD-RUSH-SETUP-GUIDE.html.

Dashboard

The companion dashboard is a static Vite + React app that reads jobs.json / meta.json from public/. It's password-gated via a Vite-baked env var (VITE_DASHBOARD_PASSWORD) — note this is client-side obfuscation, not real auth; fine for single-user personal use, not for anything sensitive.

The pipeline writes fresh JSON to the dashboard repo on each run and git pushes; Vercel handles the deploy.

License

MIT