Philip Kiely's book covers the components needed to run models in production: hardware, software, techniques, and infrastructure. According to the book's description, its main focus is LLMs, with some content on diffusion models.
Use this broad view to organize your studies. Group topics under hardware, software, techniques, and infrastructure, then note how each relates to running models in production. The source does not specify chapters or particular technical recommendations.
When applying the material to a real system, turn each topic into questions for your team to investigate rather than assuming the book solves specific engineering challenges. Record decisions and validate them against your workload.
If you use AI to study or apply the material, avoid submitting personal data, credentials, confidential code, or internal information without authorization. Rota Nacional detects personal data and applies the organization's policy before a model runs; that policy can replace detected data with markers, remove it, or block the request.