About this talk
This talk by Vincenzo Gambino explores the intricacies of Retrieval-Augmented Generation (RAG), particularly in the context of managing large datasets such as extensive Drupal nodes. The session covers foundational concepts of RAG and embeddings, as well as advanced techniques that improve the reliability of LLM outputs. Key topics include hybrid retrieval methods that merge semantic embeddings with traditional keyword search, smarter chunking strategies that utilize content structure, and re-ranking processes that enhance retrieval accuracy. Attendees will gain insights on how to develop a more sophisticated AI assistant that effectively utilizes their content, enriching the user experience while scaling with Drupal.
