MLOps: Leveraging DevOps to run ML systems

- AI is huge. Huge I tell you. Machine learning is taking everything over. - Okay, well that may be a bit of an exaggeration. - Chat GPT told me you were going to say that. - Okay well, while machine learning's been around for a long time, it's mostly been used by huge social media companies, and hardcore scientists, and engineers up until recently. Now, the boom in generative AI has every single company scrambling to come up with an AI ML strategy. - And of course, while you can play around with web-based AI tools, most of the ways to create business value with AI involve lots of private data, and lots of training of private machine learning models. - Okay, well that sounds like it needs infrastructure and it's a DevOps problem. - It is, but with a little bit of a unique spin on it. And thus the term MLOps was born, combining machine learning and DevOps. - Okay well, what makes MLOps different from just normal DevOps? I mean, you spin up servers, you send out jobs, you monitor them…