RRisingWave Labs·Apr 7How to Serve ML Features Under 10ms Using Materialized ViewsPre-computed materialized views turn expensive aggregations into cheap key lookups — here's how to build a low-latency feature serving layer.11 min read·risingwave.com