3 ms·

I went through it twice, once for classical ML engineering work (used to build bespoke models, not just prompt), and second time for coding.