All updates

AI-powered search

Hybrid search combining semantic vector embeddings, keyword matching, and neural reranking. Candidates find relevant jobs even with imprecise queries. "Ruby developer" surfaces "Rails engineer" results too.

AJ
By Abi Tyas Tunggal and Jack Walsh on

Search on Cavuno boards uses AI to understand what candidates mean, not just what they type.

How it works

  • Semantic vector embeddings powered by Voyage AI encode job titles, descriptions, and skills into a shared meaning space. Jobs are matched by concept, not just keywords
  • Hybrid retrieval blends dense vector search with sparse keyword matching via Qdrant. Precision of keywords with the recall of semantic understanding
  • Neural reranking re-scores the top results for relevance. The most useful jobs surface first
  • Typo tolerance: misspelled queries still return the right results. No dead-end searches
  • Faceted filtering: candidates refine results by location, salary, job type, experience level, skills, and more. Filters update counts in real time

What this changes

On a traditional job board, searching "Ruby developer" won't find a listing titled "Rails engineer," even though they're the same role. Semantic search closes that gap. Candidates find more relevant jobs, stay longer, and apply at higher rates. Better search means better retention and more revenue.