The hitchhikers practical guide to MLOps

So you find yourself in AI land because the CXO said it was important. We now have systems utilizing traditional ML systems to GenAI and everything in between. Some systems are build by the data science team in something call notebooks and some are using proprietary SaaS models running through some obscure APIs. But how do we run them? Observe them? Improve them? Understand them and debug them? Bring your towel and lets explore a practical demo of how MLOps can be implemented together with some generalized learnings about going from one model to hundreds of models in production. We will lean on everything we learn along the way from DevOps to SRE and even the Cloud Native movement. But you already know the answer. It’s 42.