Train a model that you can trust, from the business question to a handoff that another person can run. You build a ticket escalation classifier step by step, and test it honestly at every step.
Read every lesson as a guest, or sign in to enroll and track your progress.
Enroll to save the lessons that you complete. It is free.
By the end of this course, you will be able to:
A machine learning model is useful only if it helps someone make a better decision, and only if its test result is honest. Most failed models fail quietly: the score in the notebook was high because the test data leaked into training, the metric did not match the decision, or nobody checked a simpler option.
This course teaches the full path with one running example: Larkfield, a fictional online shop for home and garden products, wants to know when a support ticket arrives whether it will need a senior agent. You use the data that you prepared in the SQL course, now for about 1,000 customers over two years. You start with what machine learning is and how to judge a model honestly. Then you begin with a rule that anyone could write, train linear models, trees and ensembles, and compare them with the same honest procedure. You also learn when a simple model is the better choice.
At the end, you deliver a ticket escalation classifier: a saved pipeline that loads in a fresh process, a baseline comparison, a leakage-safe evaluation with a threshold that fits the team’s capacity, an error analysis and a model card that says what the model must not be used for.
You completed Python for Practical AI Engineering, Essential Mathematics and Statistics for AI, and SQL and Data Preparation for AI, or you have the same skills. You need a free Google or Kaggle account for the notebooks, or Python 3.12 or later on your computer. All tools are free.
Courses to take first: Python for Practical AI Engineering, Essential Mathematics and Statistics for AI, SQL and Data Preparation for AI.
People who can prepare data in Python and want to understand machine learning from the start, and then train, evaluate and explain their first real models. Fresh graduates and professionals moving into AI roles.
You can study the ideas of this course without writing code. Read modules 1–3 (about 6–7 hours), answer their knowledge checks, and get a free concepts certificate. You can come back for the practice later.
The paths for people who lead or use AI take this course this way. See the paths
Experienced AI practitioners designed this course. AI helped us draft the content; people chose the teaching approach and the examples, and reviewed every lesson. The outputs that you see are the actual results of running the code. How we make our courses
Learn what machine learning is, starting from how people learn. You see why learning from labelled examples is called supervised learning, how unsupervised learning finds groups without labels, and how to choose regression, classification, clustering or no model at all.
Turn a business need into a prediction problem. You name the decision, the target and the moment of prediction, and set a simple baseline that any model must beat.
Learn how to judge a model honestly before you build one. You keep separate data for practice, for checking and for the final exam, and you choose metrics that fit the task, including precision, recall and the trade-off between them.
Train your first models. You fit a line to daily ticket counts, turn scores into decisions, and see which problems a scikit-learn pipeline solves before you build one.
Improve a model the way practitioners do, one measured experiment at a time. You meet decision trees, see when a model is too simple or too complex, and learn why ensembles of many trees can help and what they cost.
Finish the model. You test which features help without leaking the answer, tune hyperparameters with cross-validation, pick a threshold for the team's capacity, and report the test result once.
Find where the model fails and why. You compare groups, read what the model relies on without mistaking it for cause, learn when a simple model is the better or the required choice, and explore the data with clustering.
Hand over a model that others can run and trust. You save the full pipeline with its schema and versions, write a model card, and plan the checks that it needs in production.
Seven end-to-end case studies on real public data, from the raw file to a tested result and a recommendation. Two regressions, three classifications (logistic regression, a decision tree, and a comparison of four model families) and two clusterings. Complete at least one for the certificate; each one is a portfolio piece.
Complete at least one case study for the certificate. The others are optional. What each one covers
Use every module in one practice project. You collect evidence, and check it against the course outcomes.
Realistic projects from start to finish, on real data, with every output explained. Each one makes a portfolio piece. Complete at least one for the certificate; the others are optional.
Forecast bike-share demand2 hr
Regression on hourly rentals: leaks, a time split, Poisson models and error slices.
Regression on real house sales: a log target, location, outliers and what a 13% error means.
Predict credit-card default2 hr
Logistic regression with a cost-based threshold, calibration, and fairness slices.
Rank a call list2 hr
A decision tree that ranks calls under capacity: the duration leak, pruning, lift and drift.
Choose a model for hospital readmissions2 hr
Compare logistic regression, a tree, a forest and boosting under one protocol, and choose.
Segment online-shop customers2 hr
Clustering customers on recency, frequency and spend, checked against their future purchases.
Find wholesale buyer types1 hr 40 min
Clustering skewed spending data: scaling, the silhouette trap, PCA and an existing label.
Other people teach this subject well too. We chose these to go with this course. They open on other sites.See all community picks
Loading the reviews…
Took this course? Sign in to review it. Learners can write a review after they complete 25% of the lessons.
