PMLE Automating and Orchestrating ML Pipelines Practice Question

A team is using Vertex AI Pipelines to orchestrate a training workflow. They want to ensure that the pipeline can be reproduced exactly six months later for auditing purposes. They need to capture all necessary information to rerun the pipeline and obtain identical results. (Choose two.)

⚠ Common exam trap

The trap here is assuming that caching or experiment tracking alone provides reproducibility, but they only log or reuse outputs without guaranteeing the exact code and data are preserved.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Use a fixed pipeline template with pinned component versions and container image digests.

To reproduce a pipeline run exactly, you must pin the pipeline template and container images to immutable versions, and store the exact parameters and input artifacts in a versioned location. These two practices ensure that both the code and data are preserved, allowing an identical rerun. Other options like caching or experiment tracking do not provide the necessary immutability.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Rely on the default caching behavior of Vertex AI Pipelines to reuse previous step outputs.
Why it's wrong here
Caching can speed up reruns by reusing outputs, but it does not guarantee reproducibility because the cache might be evicted or the component might be re-executed if inputs change. Caching is not a substitute for capturing the exact code, parameters, and data. It also does not help if you need to audit the original run's inputs and outputs.
✗
Use the same pipeline name and run it again with the same parameters.
Why it's wrong here
Using the same pipeline name does not guarantee reproducibility because the pipeline template might have changed, container images might have been updated, or input data might have been modified. Without pinning versions and capturing inputs, rerunning with the same parameters could yield different results. The pipeline name is just an identifier and does not enforce immutability.
✓
Use a fixed pipeline template with pinned component versions and container image digests.
Why this is correct
Pinning component versions and container image digests ensures that the exact same code and dependencies are used when the pipeline is rerun. This is critical for reproducibility because container images can be updated, and using digests guarantees immutability. Pinned versions also prevent unexpected changes in component behavior.
✓
Store the pipeline parameters and input artifacts in a versioned location, such as a Cloud Storage bucket with versioning enabled.
Why this is correct
Storing parameters and input artifacts in a versioned location allows you to retrieve the exact inputs used in the original run. Cloud Storage versioning preserves previous versions of objects, so even if files are overwritten, the original data remains accessible. This is essential for reproducing the pipeline run.
✗
Enable Vertex AI Experiments to log metrics and parameters for each run.
Why it's wrong here
Vertex AI Experiments logs metrics and parameters, which is useful for tracking, but it does not capture the full pipeline definition, container images, or input data. It provides a record of what happened, but not the means to rerun the pipeline identically. Reproducibility requires capturing the pipeline template, parameters, and artifacts.

Go deeper

PMLE Automating and Orchestrating ML Pipelines Practice Question

Why each option matters

Option-by-option breakdown

Related to this question