ML pipelines
Aidin Pipelines is Aidin’s managed pipeline engine for repeatable, multi-step ML workflows: compile a pipeline from Python (or upload a YAML DAG), run it, and compare runs, metrics, and artifacts.
What you get
Section titled “What you get”- Pipeline runs and experiments — every execution is tracked with its parameters, status, and output metrics; runs are grouped into experiments for comparison.
- Artifact storage — run outputs are stored per-project in the platform’s artifact storage.
- Two authoring paths — Python SDK compilation or uploaded pipeline YAML.
Writing pipelines
Section titled “Writing pipelines”The Aidin Python SDK provides pipeline authoring helpers (@pipeline / @component style wrappers) on top of KFP — see the Python SDK documentation. Pipelines you compile run in your project’s namespace under your project quota.
Scheduling and resources
Section titled “Scheduling and resources”Pipeline steps run as workloads in your project namespace and are subject to the same quota and scheduling rules as notebooks: steps queue when quota or capacity is busy.