jobsearch
Python CLI tool that searches for jobs via Google Custom Search API with smart filtering to prioritize company career pages over aggregators.
2 commits over 1 day
What it is
A command-line job search agent that uses Google Custom Search API to find job postings across the web. It features intelligent domain classification to filter out job aggregator sites (Indeed, LinkedIn, etc.) and surface direct company career pages. Supports hybrid search modes, batch processing of multiple job titles, MySQL database storage for results tracking, and exports to JSON/CSV. Built for job seekers who want higher-quality, direct-to-company job listings rather than wading through aggregator spam.
Why I built it
Likely built to solve a personal job-search frustration: generic Google searches return 80% job aggregators and only 20% actual company career pages. The comprehensive filtering system and domain classifier suggest the developer wanted direct access to company job postings without the noise. The hybrid search architecture and database integration indicate an evolution toward a more sophisticated job-hunting toolkit.
Tech stack
Python 3.7+, Google Custom Search API, MySQL for result persistence, pandas for data handling, OpenAI SDK (likely for query enhancement or text processing). Rich CLI with progress indicators, CSV/JSON export, environment-based config (.env), and comprehensive testing suite.
Timeline & effort
Last active Apr 14, 2025 — 2 commits across 1 active day (spanning 0 days).
Very lightweight commit history—just 2 commits on a single day (August 14, 2025)—but the codebase and documentation tell a deeper story. The repo contains 30+ Python modules, extensive markdown planning docs (FILTERING_IMPLEMENTATION_SUMMARY, FILTER_APPROACH, HYBRID_SEARCH_APPROACH), a 70+ domain classifier, and a full test suite. This suggests either bulk-committed work from local development or agentic/AI-assisted rapid scaffolding in a single session.
Did it turn out useful?
The implementation is complete and production-ready: CLI with multiple search modes, database integration, batch processing, analytics module, comprehensive filtering (70+ aggregator domains identified), and detailed setup docs. The FILTERING_IMPLEMENTATION_SUMMARY shows before/after metrics (80% aggregators → 100% company pages) and real-world examples. The hybrid search architecture, quality scoring, and deduplication logic indicate this was genuinely used and refined for actual job searching, not just prototyped.