Exercises + milestone requirements owned by this stage
Evidence and gate
0 · 0/9
Artifacts · checked requirements
Why it matters
Purpose of this stage
Train/evaluate conventional ML
Subjects
6 subjects · 49 canonical topics
Each subject states what you should be able to do, an order to work through it in, what to watch for, and the questions a reviewer should be able to ask you at the end. None of it is tracked or scored.
How to work through this subject
An order that works, not a checklist. Nothing here is tracked — the steps sequence this subject’s own topics and add none.
Write the request down as a decision someone makes, then name the prediction target that would support it.
Decide whether this is supervised or unsupervised, classification or regression, and say what in the data settles it.
List the features you could actually have at prediction time, and cross off the ones you could not.
Build the simplest baseline that could work — a constant, a rule, a single feature — and score it.
Write down what a false positive costs and what a false negative costs, in the language of the business.
Finish by stating the case in which the honest answer is that this does not need machine learning.
What to watch for
If you cannot state the baseline, you cannot claim the model helped.
The cost of a false positive and a false negative are rarely equal. Find out which one hurts before choosing a metric.
The target has to be knowable at training time and unknown at prediction time. Anything else is either impossible or leakage.
Data availability decides more projects than model choice does. Check what you can actually get before designing around it.
Framing is where most projects go wrong, and it is the cheapest place to fix them.
Curriculum topics
prediction target
The thing the model predicts. It must be knowable at training time and unknown at prediction time.
features
The inputs the model uses to make its prediction.
supervised vs unsupervised
Supervised learning has labelled answers to learn from; unsupervised finds structure without them.
classification vs regression
Classification predicts a category; regression predicts a number. The choice changes every metric that follows.
baseline
The simplest reasonable predictor. Without one you cannot claim a model helped.
data availability
Whether the data you need exists, in sufficient volume and quality, at the moment prediction happens.
business cost of errors
What a false positive costs versus a false negative. They are rarely equal, and this decides your metric.
Resources for this subject
PRIMARYMachine Learning Crash Course Google The teaching path for classical ML: framing, splitting, models, preprocessing and evaluation, with interactive exercises. Work it alongside your own dataset rather than only its examples.Open
Need another explanation?1 more reference
REFERENCERules of Machine Learning Google 43 rules from production practice, starting with 'do not use ML if you do not have to' and 'make your first model simple'. Directly reinforces the baseline discipline.Open
A reviewer should be able to answer these
Answer from memory, out loud, before moving on. These are not tracked and do not affect the gate — they tell you whether you actually understand this subject.
What is your prediction target, and could it be known at prediction time?
What is the simplest baseline for this problem, and what does it score?
Is this supervised or unsupervised, and what does the data tell you?
What is the business cost of a false positive versus a false negative here?
Why might the right answer be 'this does not need machine learning'?
How do you decide whether a problem is classification or regression?
Which of your features would be unavailable at prediction time, and how did you check?
Exercises
Demonstrated practice · 35% base weight
This stage defines implementation through its owned project milestone rather than a separate exercise list.
Resource status is a convenience marker only. It cannot change competencies, gates, or stage completion — in any tier below.
Start here
PRIMARY
Machine Learning Crash Course
Google
The teaching path for classical ML: framing, splitting, models, preprocessing and evaluation, with interactive exercises. Work it alongside your own dataset rather than only its examples.
REFERENCEscikit-learn scikit-learnThe library you will implement in, plus a user guide that explains the methods behind it. The pipeline and model-evaluation sections are the ones that prevent leakage rather than merely describing it.Two source links: Stage 8 links getting_started.html, the resource library links user_guide.html. Both preserved.Open Open
OPTIONALMachine Learning Specialization DeepLearning.AIAn optional longer course covering this stage's material with more mathematical depth. Choose it over the crash course if you want the derivations; do not run both in parallel.Open
Need another explanation?5 supplementary resources · not part of the canonical library
Use these for a demonstrated gap, not as a parallel course. They are authored additions and can never be the primary resource for a subject.
REFERENCERules of Machine Learning Google43 rules from production practice, starting with 'do not use ML if you do not have to' and 'make your first model simple'. Directly reinforces the baseline discipline.Open
REFERENCECommon pitfalls and recommended practices scikit-learnThe canonical writeup of data leakage through preprocessing, with code. Read before building the Stage 8 pipeline, not after.Open
REFERENCECross-validation: evaluating estimator performance scikit-learnCovers grouped and time-series splitting, which is where a naive random split leaks without any obvious symptom.Open
REFERENCEClassification: Accuracy, precision, recall Google MLCCWorked treatment of threshold choice and the precision/recall trade-off — the material behind the gate's demand that you reject a bad metric.Open
OPTIONALInterpretable Machine Learning Christoph MolnarFree textbook on explaining model behaviour. Useful for the slice and error analysis the Stage 8 project requires.Open
Stage gate
Readiness assessment
Not ready
Stage-specific requirements
Answer each prompt without notes. Expand a question only when you are ready to check your answer.
0/1
Reject a bad metric or leaky experiment even if its reported score is high
Answer
A high score is a claim, not evidence. Check the class balance before accepting accuracy, check whether the metric matches the decision and the cost of each error type, and check the experiment for leakage: preprocessing fitted before splitting, a random split over grouped or time-ordered rows, or a feature that could only be known after the fact. A result you cannot explain mechanically should be investigated, not published.
Universal requirements
The eight conditions 01_LEARNING_RULES.md applies to every stage. Expand one to see what it actually asks for.
0/8
Core concepts can be explained from memory.
What this asks for
Every subject in this stage has review questions. The condition is met when you can answer them out loud, without notes and without looking anything up, in your own words rather than a recited definition. An explanation you can only give with the material open is recognition, not recall.
Required exercises are complete.
What this asks for
Complete means each exercise satisfies its stated requirements and its done-when criteria, not that you attempted it. Where an exercise says you should be able to explain or defend something, that explanation is part of completing it. A stage whose source lists no exercises meets this through its owned project milestone instead.
Stage project works.
What this asks for
Works means it runs end to end from a clean checkout, on the inputs it was built for, with the failure cases it claims to handle actually handled. It is not a notebook that produced a result once. Someone else following your README should be able to run it without asking you a question.
Tests or verification exist.
What this asks for
There is something that would fail if the behaviour regressed, and you have watched it fail. That can be automated tests or a written verification procedure with recorded results, depending on what the stage produces. A test that has only ever passed is evidence of nothing.
README explains decisions.
What this asks for
The README says what the project is, how to install, run and test it, and why it is built the way it is: the choices you made, the alternatives you rejected, and what you deliberately left out. A list of commands with no reasoning does not meet this condition.
You can reproduce the core implementation without copying.
What this asks for
Delete the central piece and write it again from an empty file, without your previous version, a tutorial, or generated code in front of you. It does not have to be identical, it has to work and you have to be able to explain every line. This is the condition that separates having finished from being able.
You pass the stage self-assessment at >= 80%.
What this asks for
Record a gate exam and score yourself honestly against the dimensions the roadmap defines: recall, implementation, debugging, testing and design judgment. Elapsed time, hours watched and completed resources count for nothing here. Checking the boxes without a recorded exam does not open the gate.
Claude/ChatGPT reviewer finds no foundational blocker.
What this asks for
Ask an AI reviewer to examine the work against this stage's concepts and report foundational misunderstandings, not style. A foundational blocker is a misconception that later stages would build on, such as leakage, a broken abstraction, or an explanation that is confidently wrong. Nothing in that class may remain open when you pass.
Recorded exam
No exam recorded. Checking every requirement is not sufficient.