Searching for a job today is surprisingly painful — noisy job boards, irrelevant recommendations and algorithms that try to guess what you want and usually fail.
So I decided to experiment with a different approach and built a tool that uses a local LLM to search and filter job postings.
Meet JobStalker — an open-source Python project that:
• parses job listings from multiple sources
• summarizes and analyzes descriptions using your local LLM (Ollama, LM Studio etc.)
• ranks vacancies by relevance
• works fully offline
• keeps all your data (resume, preferences) on your machine
GitHub: https://github.com/10sorry/JobStalker
🧠 Why local LLM?
Because job search is extremely personal. I don’t want my job queries, resume or skills profile sent to remote servers.
⚙️ Tech stack
• Python
• Async pipelines
• Integration with local models
• Extensible architecture for custom ranking logic
🧩 What’s next
I’m planning to add:
• support for more platforms
• UI for browsing results
• multi-model pipeline
• custom rule-based filters
If you want to try it or contribute — I’d love your feedback.
What features would you like in a privacy-first job search tool?
Top comments (0)