This is 100% my experience. I got hired as a ML engineer to bring a data scientists models into production. I did the same as you by tearing the whole thing apa

This is 100% my experience. I got hired as a ML engineer to bring a data scientists models into production. I did the same as you by tearing the whole thing apart and engineering it properly. I also look at the models, and oh boy... that data scientist had no idea what he was doing. Couldn't explain why he chose the model, didn't have any performance metrics (or even knew what metric to use to measure the performance) and just generally did not understand the basic concepts of his fields. I had to try really hard to drag answers out of him, but in the end I came out dissatisfied.