Structure and operations.

It's typically for trainings or workshops regarding ML.

Infrastructure is an underappreciated component in the life cycle of actually building and deploying AI for use cases, but it's one of the most important aspects.

So that's why we put it in the workshop.

Long story short, Bettina introduced me a little bit, but to be honest with you, like my slides will be a report from the trenches.

So it will be how is it that infrastructure affects your life cycle to deploy something in production that has intelligence.

I've done this over and over in Silicon Valley the last 15 years, and it's a battle-tested and experience journey.

So ML infrastructure and operations, what is it?

It's emerged as an important field for deploying models in the wild in the last few years.

We have been doing it for a while, but didn't have a name.

Essentially, it's a set of processes, infrastructure, architecture and tools that ensures reproducible, robust and observable ML life cycle and deployments in production.

So there are two key words here, life cycle.

A life cycle that your artifacts or intelligent artifacts, whether it's model or data, have to go through, and deployment.

So these are the two key things that you have to remember, deployments in production.

So why are we even talking about AI?

Because we deploy to our customers, and there is a process that goes with it.

So what is the textbook life cycle for deploying anything intelligent into your product?

Anything that has to do with machine learning or AI.

Typically at the beginning, as Simion mentioned a little, you have to do something with data.

There is always a stage that involves data, massaging data, transforming data and reaching data.