MLOps — From Notebook to Production — Schedule
Week-by-week schedule of MLOps — From Notebook to Production.
The operational schedule for MLOps — From Notebook to Production. Per-cohort dates fill in at intake; the structure below is stable across cohorts.
The single source of truth is _data/mlops.yml. Edits there flow through this page automatically.
| Week | Title | Pitch | Detail |
|---|---|---|---|
| 01 | Introduction to MLOps — From Notebook to Production | Sculley's 2015 diagram: the ML model is 5% of an ML system. This course is about the other 95%. | week 01 → |
| 02 | Environments and Dependency Management | The first thing that breaks: dependencies. The current toolchain and the patterns that survive in production. | week 02 → |
| 03 | Version Control — Git and DVC | Code is the easy part. Data and models version differently. | week 03 → |
| 04 | Experiment Tracking — MLflow, Weights & Biases | Six months from now, which hyperparameter combination produced your best result? | week 04 → |
| 05 | Data Pipelines and Feature Stores | Orchestration: the chain of transformations from raw data to model-ready features, running on a schedule, observable, idempotent. | week 05 → |
| 06 | Large-Scale Training — Distributed Training | When one GPU stops being enough: data parallelism, model parallelism, and the engineering of training at scale. | week 06 → |
| 07 | Containerization — Docker for ML | Build once, run anywhere — finally true for ML, with the right image discipline. | week 07 → |
| 08 | Model Deployment — REST APIs, FastAPI, Streamlit | The model is in production when an HTTP endpoint serves it. The patterns from prototype demos to high-traffic inference services. | week 08 → |
| 09 | CI/CD for Machine Learning | Each commit must be potentially deployable. The CI/CD/CT discipline — where the third T is *continuous training*. | week 09 → |
| 10 | Model Monitoring in Production | A deployed model does not self-regulate. It drifts, finds shortcuts, gets exposed to populations the training data never saw. | week 10 → |
| 11 | Reproducibility in Research — Standards and Best Practices | Pineau and Henderson showed in 2017 that identical RL code with different random seeds produces 3× different learning curves. Reproducibility is engineering, not virtue. | week 11 → |
| 12 | Final capstone — deploy and defend a system | Each participant brings a model from any prior course and ships it through the full MLOps pipeline. Final defense, 20 minutes. | week 12 → |
Operational notes
- Default timezone: Africa/Lagos (UTC+1). Per-cohort timing negotiated at intake.
- Lab notebooks and problem-set repos live in the cohort GitHub organization.
- The bilingual lecture notes remain the reference text.