Walter Heck
- talk
RAG: 0-100km/h in 30 mins
Learn from Walter" RAG helps to avoid hallucination because I already provide the truth and the answer, in my context. So I don't need to rely on the model hallucinating because it won't be doing that because it already has the answer."
You've barely got your mind around LLMs like ChatGPT and now you see the term RAG everywhere. What is a RAG in layman's terms, why should you care and what can and can't it do?
This talk takes the attendee from 0 to RAG in simple words, explaining some of the concepts you'll see and giving some clear examples of what makes it tick. We'll cover things like embedding models and vector databases as well as context retrievers and prompt templates. After this talk you'll have a better idea about RAG and why it's exciting (it is, I promise!)
Track Sponsors
Originally presented in the Back to Basics: AI Fundamentals track
About Walter Heck
Walter Heck is a technology enthousiast with a career spanning pretty much every role in tech across over two decades. As such he's seen a few of the technology evolutions and the advent of AI is no different. He's been CTO for a non-tech company as well as consultancy founder and engineer, giving him a good experience in explaining complicated concepts in simple ways.