Vertex AI Pipelines orchestrate ML workflows on Google Cloud using Kubeflow components. Master the vocabulary for components, artifacts, lineage tracking, pipeline parameters, and Vertex Experiments for production MLOps.
0 / 5 completed
1 / 5
A data scientist defines a Vertex AI Pipeline using the Kubeflow Pipelines SDK. What is a component in this context?
A Kubeflow component (used in Vertex AI Pipelines) is a self-contained, reusable step with explicitly typed inputs and outputs. Each component runs in its own container, isolating dependencies. Components can be shared across pipelines, promoting modularity and reproducibility in ML workflows.
2 / 5
A Vertex AI Pipeline step produces a Dataset artifact that a later step consumes. What does Vertex AI artifact lineage track?
Artifact lineage in Vertex AI Pipelines (backed by ML Metadata) tracks the full provenance graph: which pipeline run produced which artifacts, which parameters were used, and how datasets and models relate across runs. This enables reproducibility audits and debugging of model quality regressions.
3 / 5
A Vertex AI Pipeline is submitted with parameter_values={"learning_rate": 0.01}. What are pipeline parameters?
Pipeline parameters are runtime inputs defined in the pipeline function signature (typed with float, str, int, etc.). Unlike hardcoded values, parameters can differ between runs — enabling hyperparameter sweeps, environment-specific configs, and reuse of the same compiled pipeline across experiments.
4 / 5
A team uses Vertex Experiments alongside their Vertex AI Pipelines. What is the primary purpose of Vertex Experiments?
Vertex Experiments provides experiment tracking: logging metrics (accuracy, loss), parameters (learning rate, batch size), and artifacts for each run. Teams can compare runs in the console, identify the best configuration, and link experiments to the pipeline runs that produced them — forming a complete ML lifecycle audit trail.
5 / 5
A Vertex AI Pipeline component is decorated with @component(base_image="python:3.11"). What does the base_image parameter specify?
The base_image in @component specifies the Docker container image that Vertex AI Pipelines uses to run that component's code. Each component can have a different base image, allowing different Python versions or pre-installed libraries. The KFP SDK injects the component function into this container at runtime.