MLOps
Note
This section is in progress. There are no MLOps articles here yet — only this outline, plus Labs and Resources placeholders.
Why this section exists
A notebook that trains once on a laptop is not a system. MLOps is the unglamorous half: how the code is structured, how a run is repeated, how a model is evaluated after training, and how serving differs from training.
What will live here
Pages I intend to write, not pages that exist today:
- How to structure ML code (the labs index already bookmarks an external note on this; it should become a real page).
- Train versus serve.
- Evaluation in the loop, pointing at Model evaluation.
- Versioning of data, code and models, in plain language.
Nothing above is written yet.