1.5.3 · Lesson 19 of 28
Vector search
Search embeddings to find semantically relevant content.
WHAT
Search embeddings to find semantically relevant content.
AWS options
- Amazon OpenSearch Service
- Amazon Aurora / PostgreSQL with pgvector
- Amazon Bedrock Knowledge Bases
Basic retrieval
Query ↓ Query embedding ↓ Vector similarity search ↓ Top-K chunks
Important concepts
- similarity
- top-K
- filtering
- metadata
- latency
- index scale
EXAM TRAP
Vector similarity does not guarantee factual correctness.
Retrieved content must still be:
- relevant
- current
- authorized
- high quality
---
Key points
- Search embeddings to find semantically relevant content.
Exam focus
Vector similarity does not guarantee factual correctness. Retrieved content must still be: relevant current authorized high quality