AI Job Discovery System

A deterministic, read-only AI job discovery system that scrapes, scores, and emails relevant AI engineering roles.

Features

• Multi-source scraping: Hacker News, X-Ray Search (Greenhouse, Lever, Ashby, Workable, etc.)
• Web-enriched company data: Real-time verification via Tavily API (employee count, funding, AI-native status)
• Deterministic scoring: LLM for parsing only, all scoring is rule-based
• Plain-text email reports: Top 20 ranked jobs with explanations

Setup

# Install dependencies
pip install -r requirements.txt

# Set environment variables (or use .env file)
export GROQ_API_KEY="your_key"
export SMTP_EMAIL="your_email@gmail.com"
export SMTP_PASSWORD="your_app_password"
export TAVILY_API_KEY="your_tavily_key"  # Optional but recommended

# Run the discovery pipeline
python -m src.main

Configuration

• config/profile.yaml: Candidate profile (skills, experience)
• config/settings.yaml: Pipeline settings (thresholds, sources, Tavily API)

New: Company Enrichment

Enable web search for verified company data (see TAVILY_INTEGRATION.md):

# config/settings.yaml
tavily:
  enabled: true
  enrich_companies: true
  max_enrichment_jobs: 30

Architecture

src/
├── main.py          # Entry point
├── models.py        # Job & Config models
├── scrapers/        # Data acquisition
├── scoring/         # Deterministic scoring
├── emailer.py       # Report generation
└── utils/           # Helpers

Scheduling

GitHub Actions runs this twice weekly (Mon & Thu at 9:00 UTC).