41 Machine learning words an engineer already knows

What people usually ask

Explain the basics of machine learning to me. I'm an analytical chemist, not a programmer.

What comes back

A glossary of twenty terms in textbook order, each defined with other terms from the same glossary, and nothing that ties any of them to work already done at a bench.

Ask this instead

I run calibration curves on an ion chromatograph: six standards, a straight line fit, then a check standard at mid range that must read within 5 percent before any sample goes on. Explain these words using only that example: sample, feature, target, training set, test set, cross validation, hyperparameter, overfitting, baseline, scaling, pipeline, leakage, residual.
One sentence each. Each sentence says what the word is in the calibration curve (the check standard is the test set, and so on). After each, in brackets, the scikit-learn name I would type: the class, the function or the argument.
At the end, three lines on where the comparison stops: what a model with ten inputs has to deal with that a one variable calibration never does.
No other examples and no list of algorithms.

Attach: Nothing.

Why it works: A word tied to a procedure already done by hand stays learned, and the scikit-learn name beside it is the bridge to the first notebook.