Yeah in general I find that cleaning the data, doing exploratory analysis, doing sensible feature engineering, etc. are much more involved tasks than running ML

Yeah in general I find that cleaning the data, doing exploratory analysis, doing sensible feature engineering, etc. are much more involved tasks than running ML algorithm XYZ.