Introduction to enterprise AI operations

- Hi, my name is Noah Gift and I'm the founder of Pragmatic AI Labs. I also lecture at some of the top universities in the world. In this particular course, we're going to talk about Enterprise Generative AI. And what's very interesting about the Enterprise versus let's say, Consumer-Facing Generative AI is you have to be much more careful. And in particular, when you think about some of the features like Generative AI, like Bedrock, et cetera, one of the things that is a plus is that they're non-deterministic, right? They can come up with interesting concepts and look at statistical patterns. But one of the downsides is that they can do the same thing, right? So, they can come up with things that are hallucinations. You also have to be very careful about trusting the kinds of outputs. So, for an enterprise way of thinking about things, it's important to go back to the roots. And what are the roots? Well, the roots are, we have things like DevOps, which is continuous improvement. We have things like MLOps, which is continuous improvement of models. So, similarly with AI or AIOps, the idea here is to incrementally make things better, but also to have a deep distrust of anything that isn't automated. Also have a deep distrust that things are going to work exactly the way you had planned them to work. So, relying on traditional tools like CI/CD, linting tools, also looking at some of the new techniques like RAG, these are some of the approaches for the enterprise that are important to consider. So in this course, we talk through many of those different techniques, including how to use prototyping Bedrock in the console, how to use it in the CLI, how to think about the RAG-based interfaces, and also how to walk through and use some of the core features of AWS, like AWS CloudShell. All right, we have a lot to cover, so let's go ahead and get started.