Machine learning foundations

About this course

A course for anyone who wants to understand what actually happens inside a machine learning model instead of just calling a ready-made AI API. It covers the full path from framing a problem and its data to a trained model: classification, regression, linear models, trees and ensembles, overfitting, and how not to fool yourself when honestly evaluating result quality.

Lessons move from problem types — 'supervised, unsupervised, reinforcement' — and metrics to linear models and trees, then to data leakage, cross-validation, and the course's central trap, overfitting: 'Bias, variance, and learning curves,' 'Regularization,' 'Tuning hyperparameters without fooling yourself.' All four course projects build on one dataset of your own — from a baseline model and an evaluation protocol to a final diagnostic report.

The course suits people who already write code or are ready to start — basic Python and some familiarity with data are needed, deep math isn't. It's a good fit for analysts and developers who want to move from using ready-made AI tools to understanding how models actually learn and why they get things wrong.

What you'll learn

• Frame a problem as classification, regression, or an unsupervised task
• Build an honest baseline before a complex model
• Split data and run cross-validation without leaking information
• Diagnose overfitting from learning curves and variance
• Tune hyperparameters without fooling yourself about the result
• Encode features, scale them, and handle missing values
• Evaluate a model with classification and regression metrics
• Interpret a model's output and understand its limits

What you'll end up with

• A baseline modeling report for your own task
• An evaluation protocol another person could reproduce your numbers from
• A diagnostic report on overfitting and tuning
• A final project writeup ready to hand off, plus a certificate on paid plans

This course doesn't cover deep learning or neural networks — it's classical machine learning fundamentals, the groundwork before more advanced topics.