For analysts, engineers and technical managers who need to frame, evaluate and discuss machine learning projects competently, with hands-on scikit-learn examples rather than theory for its own sake.
The curriculum is being prepared and reviewed. Enrollment is not open yet.
About this course
Generative AI gets the headlines, but most production predictions — churn, fraud, demand, risk scores — still come from classical machine learning, and the same concepts underpin how large models are trained and evaluated. This course gives practitioners the vocabulary and judgment to frame an ML problem, build a sound baseline and tell a trustworthy model from a misleading one.
You will work through the full loop in scikit-learn: turning a business question into a supervised or unsupervised task, preparing features with pipelines, splitting data with cross-validation and time-aware splits, training linear models, decision trees and gradient-boosted ensembles (XGBoost and LightGBM), and evaluating them with the right metric for the cost of errors. You will learn to spot overfitting, data leakage and class imbalance, interpret models with permutation importance and SHAP values, and decide when a neural network, a pretrained embedding model or an LLM is the better choice. The course ends with an automated exam and a Wirewalk Institute certificate of completion and assessment.
What you will be able to do
Frame a business question as a supervised, unsupervised or ranking problem with a clear target, unit of prediction and success metric
Build reproducible preprocessing and modeling pipelines in scikit-learn
Split data correctly with holdout sets, k-fold cross-validation and time-based splits, and detect data leakage
Train and compare linear models, decision trees and gradient-boosted ensembles against a simple baseline
Select and interpret evaluation metrics, including precision, recall, ROC AUC, PR AUC, calibration and regression errors, based on the cost of mistakes
Diagnose overfitting and underfitting and tune hyperparameters without contaminating the test set
Explain model behavior with permutation importance and SHAP, and recognize when deep learning or a foundation model is the better tool
Syllabus
6 modules · 20 lessons · about 12 hours
1
Module 1: Framing Machine Learning Problems
Turn a business question into a well-posed learning task.3 lessons
Supervised, Unsupervised and Ranking Problems
Choosing the Target, the Prediction Unit and the Decision It Supports
Baselines: The Model to Beat Before You Build Anything
2
Module 2: Data, Features and Honest Splits
Prepare data so that evaluation results mean something.3 lessons
Feature Engineering and scikit-learn Pipelines
Train, Validation and Test Sets; k-Fold and Time-Series Splits
Data Leakage: How It Happens and How to Catch It
3
Module 3: The Core Model Families
What each family assumes and where it shines.4 lessons
Linear and Logistic Regression with Regularization
Decision Trees and Random Forests
Gradient Boosting with XGBoost and LightGBM
Clustering and Dimensionality Reduction: k-Means, PCA and UMAP
4
Module 4: Evaluating Models Properly
Pick metrics that reflect the real cost of errors.4 lessons
Confusion Matrices, Precision, Recall and Thresholds
ROC AUC, PR AUC and Probability Calibration
Regression Metrics and Error Analysis by Segment
Class Imbalance: Resampling, Weights and Better Metrics
5
Module 5: Generalization and Tuning
Make models that hold up on data they have not seen.3 lessons
Bias, Variance, Overfitting and Learning Curves
Hyperparameter Search Without Fooling Yourself
Interpreting Models: Permutation Importance and SHAP
6
Module 6: From Classical ML to Deep Learning and Foundation Models
Know where classical methods end and neural approaches begin.3 lessons
Neural Networks as Learned Feature Extractors
Embeddings and Pretrained Models as Features
Choosing Between Gradient Boosting, a Neural Network and an LLM
Exam and certificate
The exam unlocks once you have completed every lesson. It draws 20 questions from a larger bank, spread across the modules, with the answer options shuffled for each attempt. You have 40 minutes and need 80% to pass. After a failed attempt you can try again after 24 hours, up to 3 attempts in total. Feedback is given per module, never as the answer key.
Passing earns a certificate of completion and assessment from the Wirewalk Institute of AI with a unique ID that anyone can check at institute.wirewalk.com/verify. It is not an accredited qualification or a vendor certification. How certification works.
The Wirewalk Institute of AI is independent. This course teaches publicly documented products and is not affiliated with or endorsed by scikit-learn, XGBoost, LightGBM, or PyTorch. Product and company names are trademarks of their respective owners.
A free, plain-English introduction taught through one realistic company getting ready for its busiest season. Sort the kinds of AI, see why a chat assistant can sound right and be wrong, decide which tasks it fits and how to measure them, catch its three predictable failures, ask well, and decide what never goes into an AI tool.
A plain-language grounding in artificial intelligence: what AI, machine learning and generative AI actually are, how chat assistants produce text, why they sometimes get things confidently wrong, and how to use them safely for real work.
A guided tour of the machinery inside a large language model, told in plain-language analogies, tables and small runnable examples rather than equations: tokenizers, embeddings, attention, pretraining, instruction tuning, reasoning models, context windows and sampling. You leave able to explain why models behave the way they do.