What's the on Source Problem in Machine Learning?

The big on-solve problem is getting everybody to do it. Make sure that you're doing this not just at the machine learning model, but all throughout the data pipeline. And I would say these three things is going to put you at like the 90th percentile of MLOps companies. It takes some effort, but it's like the 80-20 rule in effect. Like you can get 80% of the value by doing 20% of the work.