Red Hat and Elastic Fuel Retrieval Augmented Generation for GenAI Use Cases

Red Hat, Inc., the world's leading provider of open source solutions, and Elastic (NYSE: ESTC), the search artificial intelligence (AI) company, today announced an expanded collaboration to deliver next-generation search experiences supporting retrieval augmented generation (RAG) patterns using Elasticsearch as a preferred vector database solution integrated on Red Hat OpenShift AI. With this collaboration, Red Hat and Elastic are providing enterprises with the tools they need to deliver, maintain and refine RAG solutions over time on a single, consistent platform.

As organizations face the twin demands of adding AI solutions into their operations while also minimizing risk, RAG takes center stage for integrating large language models (LLM) into business applications. RAG enables IT teams to combine the benefits of LLMs with private data stores to train models with targeted, private data without modifying the underlying model itself. Strong search retrieval is key, as prompting LLMs with the correct information using private repositories at scale can be expensive. Retrieval with role-based controls helps maintain protections around sensitive data while still using it for training general-purpose LLMs.

Red Hat OpenShift AI and Elasticsearch can help organizations get the most out of RAG at both the MLOps infrastructure and application levels. Red Hat OpenShift AI provides a trusted machine learning operations (MLOps) platform to automate, build, tune, deploy and monitor models at scale. At the same time, Elasticsearch delivers a vector database and robust hybrid search solution for scaling and extracting AI responses, with advanced search and security features to make results more applicable to end users.

Red Hat supports Elasticsearch’s tools for RAG and generative AI (GenAI) application developers using the Elasticsearch Relevance Engine^TM (ESRE^TM), which includes built-in vector search and transformer models, enabling developers to build next-generation search with proprietary enterprise data. ESRE enables organizations to create deployments that are optimized for security using their proprietary structured and unstructured data, and enables developers to build semantic search and RAG applications using a variety of third-party machine learning (ML) models, as well as ecosystem tooling from providers including Cohere, LangChain and LlamaIndex.