JobMatchAI: Knowledge Graph and Semantic Search for Job Matching
JobMatchAI, an innovative system, combines Transformer embeddings, skill knowledge graphs, and interpretable reranking to enhance the matching of job candidates. Unlike conventional keyword filters, it effectively manages skill synonyms and nonlinear career trajectories, offering detailed explanations for match scores. The system maximizes utility by considering factors such as skill alignment, experience, location, salary, and company preferences. Researchers have introduced the JobSearch-XS benchmark along with a hybrid retrieval stack that integrates BM25, knowledge graph, and semantic elements. The performance is assessed through retrieval tasks. Additionally, a demo video, a dedicated website, and an installable package are provided.
Key facts
- JobMatchAI uses Transformer embeddings, skill knowledge graphs, and interpretable reranking.
- It addresses limitations of keyword-based filters by handling skill synonyms and nonlinear careers.
- The system provides factor-wise explanations for match scores.
- Optimization covers skill fit, experience, location, salary, and company preferences.
- JobSearch-XS benchmark is released for evaluation.
- Hybrid retrieval stack includes BM25, knowledge graph, and semantic components.
- Performance assessed on retrieval tasks.
- Demo video, hosted website, and installable package are available.
Entities
—