A job search engine that processes natural-language queries, generates a structured plan, and ranks results based on relevance. It uses Vespa for retrieval, applies query understanding, and employs reranking and match explanations to categorize matches as strong, partial, or weak.
From the README
Agentic Job Search and Ranking Engine An agentic job-search system that combines Vespa retrieval with query understanding, plan-aware candidate filtering, job-specific reranking, match-quality classification, and search explanations. The system is designed to return relevant job opportunities while explicitly distinguishing between strong, partial, and weak matches instead of treating every retrieved document as equally suitable. Overview The search pipeline is: Main Components agent.py Builds a structured SearchPlan from a natural-language job query. The plan can contain: • role terms • skills • locations • experience requirements • entry-level / senior-level signals • employment...
Skills
- Python
- JavaScript
- CSS
- HTML
