Machine Learning Essentials
For developers and analysts who can write Python and want a solid, practical grounding in classical machine learning. You will frame problems, train regression, classification and clustering models with scikit-learn, evaluate them honestly, and serve a trained pipeline behind a small web API.
Course Content
How to frame a machine learning problem, run the workflow in an order that prevents leakage, and diagnose underfitting and overfitting. 3 lessons, about 40 minutes.
Fit linear and polynomial models, keep them stable with ridge, lasso and elastic net, and choose error metrics that match what a mistake really costs. 3 lessons, about 40 minutes.
Build logistic regression, tree, forest and nearest-neighbour classifiers, and judge them with metrics and thresholds that survive imbalanced data. 3 lessons, about 40 minutes.
Find structure without labels using k-means, hierarchical clustering, DBSCAN and PCA, and check whether what you found is real. 3 lessons, about 40 minutes.
Save a trained pipeline safely and serve it to people with Streamlit and to programs with Flask or FastAPI. 3 lessons, about 35 minutes.
Take one problem from a business question to a tested, documented model behind an API. 1 lesson, about 15 minutes.