New course with TruEra and LlamaIndex: Building and Evaluating Advanced RAG

Hi, I'm excited to introduce Building and evaluating advanced RAG or Retrieval Augmented Generation applications built in partnership with Jeremy Liu, co founder and CEO of LLAMA Index, and Anupam Datta, co founder and chief scientist of True Era. One of the most popular use cases of LLMs is to connect an LLM with an organization's own data, a method called Retrieval Augmented Generation or rag. But to actually productionize rag, there are a couple ideas that can be very helpful. First, improved retrieval techniques so that rather than getting random blocks of text, you get the most coherent context for the LLM. And second, evaluation metrics. Like many other things in AI, implementing RAG is a highly iterative process and having certain evaluation best practices will be a big help for you to iterate efficiently to a good system.

Thanks, Andrew. In this course we'll cover two advanced retrieval methods, Sentence window retrieval and auto merging retrieval that perform better compared to the baseline RAG pipeline. To make sure that your RAG pipeline is actually performing better with these advanced methods, we need to do evaluation and experiment tracking. We will cover the RAG triad, Context Relevance, groundedness and answer relevance that are effective in programmatically evaluating and comparing rags as you experiment with various methods. Let's go. We hope you enjoy the course.

Yeah, and I hope that someone will build a really valuable business using this because that would be a RAG to riches story.