Skip to main content

PostgreSQL

10 articles · written by Niraj Kumar

PostgreSQL keeps absorbing jobs that used to need separate infrastructure: vector search via pgvector, full-text search good enough to skip Elasticsearch for many workloads, row-level multi-tenancy via RLS, and durable job or agent state. These articles lean into that.

Coverage includes hybrid retrieval combining pgvector embeddings with native full-text search using Reciprocal Rank Fusion, Row-Level Security as a hard tenant boundary for AI SaaS, LangGraph checkpointers backed by Postgres, GraphRAG-style multi-hop traversal, and a direct comparison with MongoDB for teams still choosing.

The bias throughout is toward keeping things in one database until there is a measured reason not to.

Related topics

← All articles