AI/ML in Gnok: Filtered ANN, Point-in-Time Features, and Schema-Aware Autocomplete
· 6 min read
The AI/ML stack in Gnok has been usable for a while — register an
ONNX model, build an HNSW index, materialize a feature group, run
inference inline. What's new this month is a round of upgrades
across every track: vector search with WHERE clauses
no longer falls back to brute-force, point-in-time training joins
read durably from Iceberg time-travel, and the SQL editor knows
which models you've actually registered.
This post walks through the three biggest changes — filtered ANN, offline-backed feature lookups, and the schema-aware ML autocomplete — with the SQL you can run today.