Integrating Realtime User Behavioral Embeddings Into OpenSearch

Integrating Realtime User Behavioral Embeddings Into OpenSearch - Trey Grainger, Searchkernel With the rise of semantic search approaches using content-based embeddings, query intent models integrating user signals have been overlooked in recent years within search and RAG systems. This oversight is understandable, as personalized search requires additional work collecting and processing clickstream behavior versus just “picking the latest LLM” and plugging it in to encode multimodal text and images. Thankfully, OpenSearch’s addition of User Behavior Insights now gives it a unique advantage out of the box for collecting this clickstream data. In this talk, we’ll walk through practical techniques for incorporating user context and behavioral signals to enhance search relevance, moving beyond just traditional lexical and content-embedding methods. We’ll cover how to train latent behavioral embedding models using user signals, implementing real-time personalization of search experiences with appropriate contextual guardrails to prevent over-personalization, and enabling cross-modal personalized search combining content-based embeddings with behavior-based embeddings. We’ll walk through open-source code examples that work out of the box with OpenSearch.