Data science is the process of extracting meaningful insights from data through a combination of advanced data analytics and machine learning techniques. In today's data-driven world, organizations need data science to thrive and remain competitive. It empowers businesses to unlock the hidden value within their vast datasets, enabling them to make informed decisions, optimize operations, enhance customer experiences, and drive innovation.

This solution consists of Red Hat OpenShift AI running on Dell APEX Cloud Platform for Red Hat OpenShift to illustrate how Large Language Models (LLMs) and the Retrieval Augmented Generation (RAG) framework can seamlessly process, store, and deploy large-scale models efficiently and cost-effectively in the form of a digital assistant. 

Digital assistants are prevalent across a wide range of verticals and use cases. The technology is designed to assist users by answering questions and processing simple tasks. By anchoring the model with relevant documentation, answers remain up to date and contain information unique to the organization.

PDF files or web pages are divided into smaller chunks and embeddings. These embeddings are stored in a vector database like Redis. Results from the vector database are ranked and sent to the Llama 2 model when users submit a query a semantic search is performed against the vector database. The Llama 2 model answers the queries based on results from the vector database and its pretrained capabilities.

Figure 1.   Digital assistant workflow