Building Real-Time Recommendations with Snowflake Online Feature Serving — Two-Tower Networks Modern recommendation systems face a fundamental challenge: how do you deliver personalized experiences …

Snowflake Community is a platform for users of the Snowflake cloud data platform to share knowledge, ask questions, and collaborate. Readers can learn about cloud data warehousing, data analytics, and data engineering best practices. With forums, user groups, and community events, Snowflake Community provides resources for Snowflake users to connect and learn from each other.

Snowflake Community

A real-time product recommendation system combines Snowflake Online Feature Serving with Two-Tower Neural Networks to deliver low-latency personalization. The architecture uses batch-computed item embeddings and real-time user embeddings, with features served at sub-100ms latency via configurable target_lag settings. Delta embeddings with Exponential Moving Average enable instant adaptation to user clicks within sessions, while the feature store maintains freshness within seconds. The approach balances batch processing reliability with real-time responsiveness, achieving significant performance improvements over traditional batch-only systems.