Weaviate Podcast

A Vision for The Future of Vector Search

15 snips
Dec 20, 2021
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Vector Search Meets GraphQL

  • Bob van Luijt built Weaviate to let users find graph nodes by semantic descriptions rather than exact keywords.
  • He chose GraphQL for developer-friendly traversal while keeping a vector-first data model.
INSIGHT

Embeddings Bridge Schema Gaps

  • Embeddings act as a bridge to target data objects without knowing their internal schema or keywords.
  • Weaviate stores many embedding types (text, image, domain-specific) to support varied use cases.
ADVICE

Mix Neural Search With Filters

  • Combine semantic vector queries with scalar or named-entity filters to get precise answers when needed.
  • Use distance cutoffs or returned certainty scores to tune relevance and control serendipity.
Get the Snipd Podcast app to discover more snips from this episode
Get the app