> Many of the people in this latter group had no idea how to write Python outside of a notebook, generally just structured problems to fit into XGBoost, and whe

> Many of the people in this latter group had no idea how to write Python outside of a notebook, generally just structured problems to fit into XGBoost, and when not doing that tried to squeeze resume-boosting-complexity into any problem the could find. To be fair, for 90% of business problems that require ML, I’d rather take the guy who throws XGBOOST at everything instead of the one trying to be fancy with neural networks. You get an explainable output and good results without deep subject matter expertise that the person likely won’t have. It also runs at a fraction of the cost.