Drift-Adapter: A Practical Approach to Near Zero-Downtime Embedding Model Upgrades in Vector Databases — mentioned by aayush4vedi

Wed, 10 Jun 2026 13:30:58 GMT

Changing your embedding model usually doesn't need a reindex, and that surprised me enough to build around it. Two models' vector spaces are close to a rotation of each other, so you can fit one orthogonal matrix (an SVD over a few thousand paired vectors, ~15s on CPU) and apply it to your query vectors at search time. An orthogonal transform preserves cosine exactly, so nearest-neighbour order is unchanged and the old index never moves. The whole "upgrade" is a 9 MB .npy. Drift wraps that plus the unglamorous parts of running embeddings in prod: embed() dedups across runs and tracks per-run…

Paper and discussion history. A mention is not an endorsement.

Drift-Adapter: A Practical Approach to Near Zero-Downtime Embedding Model Upgrades in Vector Databases · HackerPapers discussions

Drift-Adapter: A Practical Approach to Near Zero-Downtime Embedding Model Upgrades in Vector Databases — mentioned by aayush4vedi