← Back to Journal

Practicing Supervised Learning with scikit-learn

SEPTEMBER 3, 2026 · Career

I finished DataCamp's Supervised Learning with scikit-learn course. The ideas underneath it, classification, regression, cross-validation, weren't new. I learned those in my Master's program. This was the first time I'd sat down and practiced them in scikit-learn, exercise by exercise, instead of just applying them in a bigger project.

Supervised Learning with scikit-learn course completion badge

Knowing a concept and being fluent in it aren't the same thing

I could already explain what k-fold cross-validation does, or why Lasso zeroes out coefficients instead of just shrinking them the way Ridge does. What I hadn't done in a while was sit through drills that made me apply each idea one small step at a time. It's a different kind of practice than reading about a method or dropping it into a bigger project, more like running scales instead of playing a whole piece.

The scikit-learn API needs its own practice, separate from the theory

The concepts held up fine. What felt rustier was typing fit, predict, Pipeline, and GridSearchCV without stopping to check argument order. I don't write this code every day, so the syntax slips even when the underlying idea hasn't. Going through the course exercises was really practice for that part.

What's next

This course is the first step toward DataCamp's Associate AI Engineer for Data Scientists track, which is where I'm headed next.

Career