Machine Learning in Production: From Models to Products

Now, just as we are teaching our Machine Learning in Production course for the 10th time, the corresponding book is finally released in its hardcover version.

The hardcover book can be ordered from MIT Press or wherever books are sold. An online version of the book is and will remain available under a creative commons license at https://mlip-cmu.github.io/book/

Older versions of the chapter are still here on Medium, though we recommend to head over to https://mlip-cmu.github.io/book/ for the final version of the book.

Part 1: Machine Learning in Production: Going Beyond the Model

• Introduction (old)
• From models to systems (old)
• Machine learning for software engineers, in a nutshell (old)

Part 2: Requirements Engineering

• When to use machine learning (old)
• Setting and measuring goals (old)
• Gathering requirements (old)
• Planning for mistakes (old)

Part 3: Architecture and design

• Thinking like a software architect (old)
• Quality attributes of ML components (old)
• Deploying a model (old)
• Automating the ML pipeline (old)
• Scaling the system (old)
• Planning for operations (old)

Part 4: Quality assurance

• Quality assurance basics (old)
• Model quality (old)
• Data quality (old)
• (old: Pipeline quality)
• System quality (old)
• Testing and experimenting in production (old)

Part 5: Process and Teams

• Data science and software engineering process models (old)
• Interdisciplinary teams (old)
• Technical debt (old)

Part 6: Responsible ML Engineering

• Responsible Engineering (old)
• Versioning, provenance, and reproducibility (old)
• Explainability (old)
• Fairness (old)
• Safety (old)
• Security and privacy (old)
• Transparency and accountability (old)

All chapters are released under Creative Commons 4.0 BY-NC-ND license.