These notes represent my own study jottings and learning synthesis, not an authoritative or foolproof standard. If you find them helpful, feel free to use them. If you spot anything that could be clearer or have suggestions for additional concepts to include, issues and pull requests are warmly welcome.
Machine Learning Workflow
End-to-end lifecycle and phases of an ML project
02. Machine Learning Workflow
The end-to-end lifecycle of an ML project, from framing a decision to evaluating whether the model is worth deploying.
A model is one artifact inside a larger process. That process starts with a decision someone needs to make, passes through data collection and cleaning, and ends with an evaluation that either supports deployment or sends the work back for revision. Most of the leverage, and most of the failures, live in the parts that are not model training.
This chapter follows the workflow in order. Each stage produces something the next stage depends on, so a weak definition early on tends to surface as a mysterious evaluation problem later.