Deployment & Scaling · Practitioner
Vector Storage on AWS: S3 Vectors, OpenSearch Serverless and Aurora pgvector
Three genuinely different answers to 'where do the embeddings live', separated by an order of magnitude in cost and latency — pick on query pattern, not popularity.
- 50 min lesson
- intermediate level
- AWS AI Track
- Narrated presentation
What you’ll learn
- Vector buckets, vector indexes and the s3vectors IAM namespace
- S3 Vectors latency profile and infrequent-query positioning
- OpenSearch Serverless minimum capacity charge and hybrid search
- Aurora pgvector for SQL-filtered, transactionally-consistent vectors
- Filterable vs non-filterable metadata and size limits
- Exporting an S3 vector index snapshot to OpenSearch Serverless
Inside this lesson
A narrated, slide-by-slide presentation with 1 prediction exercise, 1 knowledge check, 4 videos. You answer as you go, and the platform checks your work.