JobHunt Engine
LiveJob hunting is a daily grind of searching, filtering and rewriting the same résumé. This agent does the grind on a schedule and leaves the decisions to a person.
- Platform
- AI agents · Python · Discord
- Built with
- OpenClaw agents
- Claude
- Grok
- Local LLM (Qwen via Ollama)
- Python
- Discord
How a day runs
Scan
Every morning it pulls new postings from job boards and company career pages and skips anything already seen.
Score
Each role is scored against target lanes, the pay floor, location rules and industry boosts.
Research and tailor
For strong matches it researches the company and builds a tailored résumé from fixed, real experience bullets.
Digest
A numbered list goes to Discord. Replying "apply 1, 3, 5" approves those applications.
Submit
Only approved applications go out. The agent never contacts an employer on its own.
What it does
- Scans job boards and company career pages every morning and skips anything it has already seen.
- Scores each role against target lanes, pay floor, location rules and industry boosts, and filters out poor fits.
- Researches the company and hiring team, then builds a tailored résumé from fixed, real experience bullets. Only the summary and emphasis change.
- Posts a numbered digest to Discord. A reply like "apply 1, 3, 5" approves the batch.
Design choices
- Human approval before any application or outreach goes out. The agent never contacts employers on its own.
- Model routing by task: a local model for high-volume structured work, cloud models for research and writing.
- Saves an estimated 10–20 hours a week compared with searching by hand.