Machine Learning
A full walking-audio course on machine learning foundations: ideas, workflows, and judgement for technical professionals. Concepts and intuition first — not library tutorials or maths derivations.
-
Lesson 1 ~10 min
What machine learning is — and when classical software is enough
Explain machine learning in plain language and decide when a rules-based program is a better fit.
-
Lesson 2 ~10 min
The main families — supervised, unsupervised, and reinforcement
Distinguish supervised, unsupervised, and reinforcement learning with everyday examples.
-
Lesson 3 ~10 min
Data, features, and labels — the raw material of learning
Describe how raw information becomes features and labels a model can learn from.
-
Lesson 4 ~10 min
Train, validation, and test — honest evaluation
Explain why data is split and what each split is for.
-
Lesson 5 ~10 min
Loss, error, and what 'learning' actually optimises
Explain loss as a score of how wrong the model is, and why optimisation matters.
-
Lesson 6 ~10 min
Regression — predicting numbers
Describe regression problems and how to judge if a numeric predictor is useful.
-
Lesson 7 ~10 min
Classification — predicting categories
Describe classification problems and the difference between scores, thresholds, and hard labels.
-
Lesson 8 ~10 min
Overfitting, underfitting, and regularisation
Explain the bias-variance trade-off in everyday language and what regularisation is trying to do.
-
Lesson 9 ~10 min
Trees, forests, and ensemble intuition
Explain decision trees and why combining many models often works better.
-
Lesson 10 ~10 min
Neural networks — the big idea without the scare factor
Describe a neural network as layered transformers of features, and when deep learning is worth it.
-
Lesson 11 ~10 min
The practical ML workflow — from question to monitored model
Walk through an end-to-end ML project lifecycle a team can actually run.
-
Lesson 12 ~10 min
Ethics, bias, and knowing when to stop
Spot common failure modes — biased data, feedback loops, and over-automation — and ask better questions before shipping.