[tomcat27]

Yeah, tooling is the heavy work. Most of the times a simple KNN model baseline is good enough starting point.

People are impatient to do math, they'd rather want to try 100 things at once, see what works. So I am not surprised they don't care in the interviews.

Only when someone writes a fancy engineering blog post or paper, weird equations appear out of nowhere..

This is so much visible in deep learning papers. The equations are there just to impress.