Get the latest tech news
Sharding Pgvector
pgvector Mar 25th, 2025 Lev Kokotov If you find yourself working with embeddings, you’ve shopped around for a vector database. pgvector is a great option if you’re using Postgres already.
Once you reach a certain scale (about a million arrays), building indices starts taking a long time. PgDog supports parallel cross-shard queries, so a map-reduce across your n vector indexes (or even just table scans, for perfect recall) is feasible. For this to work, you’d need to increase your original K-means centroids number to something high enough that will split your dataset just right.
Or read this on Hacker News