Prompt Engineering Guide creator
Elvis Saravia
Profile
Elvis Saravia is the person behind the Prompt Engineering Guide — the open-source resource at github.com/dair-ai/Prompt-Engineering-Guide that has served as the de facto reference for developers learning how to talk to LLMs. It started in December 2022, right after ChatGPT landed, when almost nobody had a coherent name for what they were doing with these models. Elvis did, and he wrote it down. The repo has tens of thousands of stars and gets cited constantly in courses, internal docs, and onboarding guides across the industry.
He has a PhD in computer science with a focus on NLP and language models, and before going independent he was at Meta AI as a technical product marketing manager supporting teams like FAIR, PyTorch, and Papers with Code. He’s a co-creator of Galactica, Meta’s science-focused LLM. Earlier, he was an education architect at Elastic building technical curriculum — which explains a lot about why his materials are so cleanly structured.
Today he runs DAIR.AI as co-founder and lead researcher, where he splits time between original research on LLM evaluation and agents, and teaching. His Maven cohort courses (Advanced Prompt Engineering for LLMs, LLMs for Everyone) run regularly and sell out. His Substack and X account (@omarsar0) are also where a lot of developers first hear about new papers — he has a knack for surfacing the useful ones without the hype spiral.
What matters for someone building with AI: Elvis’s work is the cleanest on-ramp that exists for getting from “I’ve used ChatGPT” to “I can engineer a reliable prompt pipeline.” The guide doesn’t push a product or a framework. It catalogs techniques — few-shot, chain-of-thought, ReAct, self-consistency, tree-of-thoughts — with working examples and citations to the original papers. If you want to understand what Jason Wei meant by CoT or why ReAct became the default agent pattern, this is where you read it.