Vector Databases: Using Vector Search with Graph and Key-Value Queries
summary
Aerospike embeds vectors alongside key‑value records and graph data, implements HNSW index partitioning across a distributed query layer with caching and replica merges, and preserves high write throughput while keeping reads fresh. A RAG pipeline demo shows chunked docs → embeddings stored in Aerospike and a retrieval step that returns results in about 10 ms search. A retail demo combines vector similarity, graph “bought together” traversals and secondary filters to refine results.
AI-written summary · about
chapters (8)
- 0:00Vector databases intro
- 3:42What vectors are
- 10:06How approximate nearest neighbor search works
- 14:20Aerospike vector search
- 18:54Multimodel benefits
- 23:30Filtering and latency
- 28:30Demo RAG pipeline
- 30:37Demo retail search
related talks
01h 23mVector Databases Demystified: Discover how they work
39mWhat's in a Vector Database?
33mVector Databases for GenAI applications
27mUnified AI & Relational Power: Scalable Distributed Vector Search with
06mWhat makes vector databases essential for RAG and AI apps?
32mDistributed Vector Databases - What, Why, and How
08mBuilding a Real-time Data Future
36mUnleashing the Power of Vector Databases for AI-Driven Applications