ML Systems Taxonomy
All four volumes Draft
The ML Systems Taxonomy
A shared map of what an ML systems engineer needs to know. It is organized as the lifecycle that builds a system, the stack it runs on, and the qualities it is judged by, on shared foundations. Each subject has a code you can cite, and each concept has a page showing what it affects and where MLSysBook teaches it.
About this taxonomy: scope, method, tagging, and known gaps
No published, community-endorsed body of knowledge for ML systems existed when this draft was written. Courses split between a compute-stack tradition (deep learning systems) and a production tradition (ML systems design and MLOps). This taxonomy joins the two and adds what both tend to leave out: datacenter infrastructure, safety and assurance, sustainability, and agentic and embodied systems.
Each concept has one home, the subject that owns its mechanism; other placements appear as related subjects. The taxonomy organizes knowledge for browsing and does not prescribe a course order or prerequisites. Relationships are editorial drafts, each with the condition under which it holds. Book links point to where MLSysBook teaches a subject; a missing link is unfinished alignment, not evidence that the books omit it.
This is an editorial draft for expert review. It is not a claim of completeness, consensus, or demonstrated learning.
Loading the taxonomy…
Download & cite
Use the codes to tag course materials, map a syllabus, classify papers or questions, or find gaps. Every file is generated from the same source as this page.
| Subjects | CSV | One row per subject with code, definition, and concepts |
| Concepts | CSV | One row per concept with definition, aliases, and subject |
| Complete taxonomy | JSON | Subjects, concepts, and typed relationships |
| SKOS vocabulary | RDF | W3C standard format for taxonomies |
| Concept graph | GraphML | For Gephi, yEd, or a graph database |
| Printable outline | Print or save the full outline |
Cite as