Vector Search at Scale — HNSW + Sharding + Filtering

HNSW = Hierarchical NSW graph

Vectors linked into a layered graph. Search descends from top (few nodes) to bottom (all nodes). O(log N) lookup.

Advertisement

Vectors linked into a layered graph. Search descends from top (few nodes) to bottom (all nodes). O(log N) lookup.

Advertisement

Beyond ~10M vectors per shard, split. Query fans out to all shards. Merge top-K globally.