Overview
Learn how Yelp incorporates LLM-powered annotations into their search system through this conference talk from Haystack EU 2024. Discover the technical implementation of review highlights annotations that explain business relevance to users, focusing on the integration of LLM expansions with existing search infrastructure. Explore the challenges faced in scaling the annotation system to full traffic capacity, managing large volumes of textual data, and adapting the system for internal RAG applications. Gain insights from Tech Lead Pallavi Patil's experience in designing backend solutions for Search Engine Results Pages (SERPs), including improvements to search map views, carousel features, and recall strategies. Understand how Yelp leverages user reviews and business information to enhance search result relevance while maintaining system efficiency.
Syllabus
Haystack EU 2024 - Pallavi Patil: Relevance Proof in Yelp Search: LLM-Powered Annotations
Taught by
OpenSource Connections