LLM Foundations: Vector Databases for Caching and Retrieval Augmented Generation (RAG)

As large language models grow in popularity, the infrastructure to be used around them also becomes vital to reduce costs, generate accurate responses, and improve efficiency. Vector databases play a vital role in several LLM use cases to help alleviate LLM shortcomings, reduce costs and latency. Knowledge of its basics and applications are vital for any engineer building applications with LLMs, and in this course, Kumaran Ponnambalam teaches you the basics of vector databases and how to use them in LLM caching and retrieval-augmented generation (RAG).

Kumaran begins with a discussion on the basics of vector databases and their applications. He then explores specialized databases for storing vectors and uses the Milvus database as the reference example, and demonstrates read and write operations with the Milvus database. Learn how to use vector databases for LLM caching, with an example use case, along with examples of RAG use cases. Finally, Kumaran concludes with a discussion on optimizing vector databases.