PMLE Automating and Orchestrating ML Pipelines Practice Test 2 — 15 Questions

Question 1

A data scientist is creating a Vertex AI pipeline using the Kubeflow Pipelines SDK v2. Which TWO statements about pipeline parameters are correct? (Choose two.)

Accepted Answer

Pipeline parameters are defined as inputs to the pipeline function decorated with @dsl.pipeline.. In the Kubeflow Pipelines SDK v2, pipeline parameters are explicitly defined as input arguments to the pipeline function that is decorated with @dsl.pipeline. These parameters serve as the primary mechanism for passing configuration values (e.g., model name, learning rate, number of epochs) into the pipeline at creation time and can be consumed by downstream components.

Answer

Pipeline parameters must be serialized to JSON before use.

Answer

Pipeline parameters can only be of type str.

Answer

Pipeline parameters can be used to pass large datasets between components.

Question 2

A data science team wants to build a Vertex AI pipeline that trains a model, evaluates it, and conditionally deploys it if the accuracy exceeds 0.9. They want to use the Kubeflow Pipelines SDK v2. Which construct allows them to conditionally execute the deployment step based on the evaluation metric?

Accepted Answer

dsl.If. In Kubeflow Pipelines SDK v2, `dsl.If` is the correct construct for conditionally executing pipeline steps based on runtime metrics or parameters. It allows you to define a condition that, when evaluated to true, triggers the deployment step only if the model accuracy exceeds 0.9. This is the standard way to implement branching logic in v2 pipelines.

Answer

dsl.Conditional

Answer

dsl.ExitHandler

Answer

dsl.Collected

Question 3

A machine learning engineer needs to schedule a Vertex AI pipeline to run daily at midnight. Which approach should they use?

Accepted Answer

Create a Cloud Function that calls the Vertex AI API, triggered by a Pub/Sub message from Cloud Scheduler.. Vertex AI Pipelines does not natively support cron scheduling. The recommended pattern is to use Cloud Scheduler to publish a message to a Pub/Sub topic at the desired time, which then triggers a Cloud Function that calls the Vertex AI API to create and run the pipeline. This decoupled architecture ensures reliable scheduling and allows for custom logic before invocation.

Answer

Use the Vertex AI Pipelines console to set a cron schedule directly on the pipeline.

Answer

Create a Cloud Build trigger that runs the pipeline on a schedule.

Answer

Use Cloud Tasks to create a recurring task that invokes the pipeline.

Question 4

You are building a Vertex AI pipeline using the KFP SDK v2. One component processes a large dataset and outputs a metrics artifact. You notice that the component is being cached even when the dataset changes, because the component code and image remain the same. How can you force the component to always re-execute when the dataset changes?

Accepted Answer

Set caching_strategy.max_cache_staleness = "0s" on the component.. In KFP SDK v2, caching is keyed on the component code, image digest, and all input values. If the dataset is passed as a URI string, changing the URI will invalidate the cache. If the dataset changes without a URI change, you can disable caching per component by setting caching_strategy.max_cache_staleness to 0 or using the disable_cache method.

Answer

Use the dsl.CacheKey annotation to explicitly set cache keys.

Answer

Change the component's image tag to :latest.

Answer

Add a random integer parameter to the component to vary the inputs.

Question 5

A team wants to implement continuous training for their ML model. The pipeline should be triggered when new training data arrives in a Cloud Storage bucket. Which combination of services should they use?

Accepted Answer

Cloud Storage → Cloud Functions → Vertex AI Pipelines. Event-driven triggers for Vertex AI pipelines are typically implemented using Cloud Storage notifications to Pub/Sub, which triggers a Cloud Function that creates a pipeline job. Cloud Build is for CI/CD, not for event-driven data triggers.

Answer

Cloud Storage → BigQuery → Vertex AI Pipelines

Answer

Cloud Storage → Cloud Build → Vertex AI Pipelines

Answer

Cloud Storage → Cloud Scheduler → Vertex AI Pipelines

Question 6

You are using KFP SDK v2 to define a pipeline. You need to pass a large dataset between components. What is the best practice for passing data?

Accepted Answer

Write the data to Cloud Storage and pass the GCS URI as an artifact.. In KFP SDK v2, passing large datasets between components is best done by writing the data to Cloud Storage and passing the GCS URI as an artifact. This approach leverages KFP's built-in artifact tracking, ensures data persistence across container restarts, and avoids memory or disk limitations of ephemeral containers. The artifact is automatically serialized and passed as an input/output parameter, enabling efficient, scalable data exchange.

Answer

Use the component's temporary directory to share data between containers.

Answer

Pass the data as a serialized Python object in memory.

Answer

Store the data in a BigQuery table and pass the table reference.

Question 7

A machine learning engineer wants to use a pre-built Google Cloud Pipeline Components (GCPC) to train a model using Vertex AI. Which component should they use?

Accepted Answer

VertexTrainJobOp. The Google Cloud Pipeline Components library includes pre-built components for various Vertex AI services. For training, the correct component is VertexTrainJobOp. VertexBatchPredictOp is for batch prediction, VertexEndpointCreateOp for deploying endpoints, and AutoMLTabularTrainingJobRunOp is specifically for AutoML tabular jobs.

Answer

AutoMLTabularTrainingJobRunOp

Answer

VertexEndpointCreateOp

Answer

VertexBatchPredictOp

Question 8

You are developing a Vertex AI pipeline that runs multiple parallel training jobs with different hyperparameters, then collects their results and selects the best model. Which KFP SDK v2 construct should you use to run the parallel training tasks?

Accepted Answer

dsl.ParallelFor. In KFP SDK v2, `dsl.ParallelFor` is used to run parallel training tasks with different hyperparameters by iterating over a list of parameters and executing a pipeline component for each item in parallel. This construct enables the pipeline to run multiple branches concurrently. After parallel execution, `dsl.Collected` can gather the outputs for aggregation, but the question asks for the construct to *run* the parallel training tasks, not to collect results.

Answer

dsl.Metrics

Answer

dsl.If

Answer

dsl.Collected

Question 9

A company uses Cloud Composer to orchestrate their ML workflows. They have an Airflow DAG that runs a Vertex AI pipeline, then a BigQuery query, then a Dataflow job. The DAG is failing because the Vertex AI pipeline takes longer than the Airflow task timeout. What is the best way to handle this?

Accepted Answer

Use the Airflow Vertex AI pipeline operator with wait_for_completion=True.. Vertex AI pipelines are asynchronous; the Airflow operator should wait for completion using a sensor or by setting the wait_for_completion parameter. Increasing the task timeout is a workaround but not best practice. Using a separate DAG for the pipeline defeats the purpose of orchestration.

Answer

Increase the Airflow task timeout to account for the maximum expected pipeline duration.

Answer

Run the pipeline synchronously by using a Kubernetes Pod operator.

Answer

Split the DAG into two DAGs: one for the pipeline, one for the rest.

Question 10

You are defining a Python function component in KFP SDK v2. Which decorator should you use?

Accepted Answer

@dsl.component. In KFP SDK v2, the `@dsl.component` decorator is used to define a Python function as a lightweight, reusable pipeline component that can be executed independently. This decorator automatically generates a containerized component from the function's signature and type annotations, enabling type-safe inputs and outputs without requiring a separate component YAML specification.

Answer

@dsl.task

Answer

@component

Answer

@dsl.pipeline

Question 11

A team wants to implement continuous delivery for their ML models. They have a pipeline that trains a model and evaluates it. If the evaluation metrics exceed a threshold, the model should be deployed to a staging endpoint, and after manual approval, to production. Which approach should they use?

Accepted Answer

Use Cloud Build to orchestrate the whole process, with a manual approval step before deploying to production.. Cloud Build supports manual approval steps via its 'approval' configuration in the build YAML, allowing the team to gate the production deployment after staging evaluation. This aligns with the requirement for continuous delivery (not deployment) where a human-in-the-loop approves the final production rollout. Vertex AI Pipelines lacks native manual approval gating, and the other options bypass the required manual approval step entirely.

Answer

Use Vertex AI Pipelines to deploy to production directly after evaluation.

Answer

Use Cloud Scheduler to trigger deployment every hour.

Answer

Use Cloud Functions to deploy after evaluation without approval.

Question 12

You are designing a Vertex AI pipeline that includes a container component. The component needs to use a custom container image that is stored in Artifact Registry. How should you specify the container image in the component definition?

Accepted Answer

Use the @dsl.container_component decorator and set the image parameter to the URI.. The `@dsl.container_component` decorator is specifically designed for defining container components in Vertex AI pipelines. The `image` parameter accepts the full Artifact Registry URI (e.g., `us-central1-docker.pkg.dev/my-project/my-repo/my-image:tag`), allowing the pipeline to pull the custom container from Artifact Registry at runtime. This decorator also requires you to define inputs and outputs explicitly, ensuring the component is properly integrated into the pipeline graph.

Answer

Use ContainerOp class from kfp.v2.dsl.

Answer

Use a placeholder in the pipeline YAML.

Answer

Use the @dsl.component decorator and set the base_image parameter.

Question 13

You have a Vertex AI pipeline that trains a model and outputs a Model artifact. You want to register this model in the Vertex AI Model Registry. Which pre-built Google Cloud Pipeline Components component should you use?

Accepted Answer

VertexModelUploadOp. The correct component is VertexModelUploadOp because it is specifically designed to upload a trained model artifact to the Vertex AI Model Registry, creating a new model version or a new model if one does not exist. This component takes the model artifact from a pipeline step and registers it, making it available for deployment or version management.

Answer

VertexEndpointDeployOp

Answer

CreateModelVersionsOp

Answer

ModelRegisterOp

Question 14

You are building a CI/CD pipeline for an ML model using Cloud Build. When code is pushed to the main branch, you want to automatically build a training image, run a Vertex AI pipeline, and if the model evaluation passes, deploy it to a staging endpoint. Which two components are essential for this CI/CD pipeline?

Accepted Answer

Vertex AI Pipelines to orchestrate training and evaluation.. Vertex AI Pipelines is the service that orchestrates the entire ML workflow, including training, evaluation, and conditional deployment logic. In a CI/CD pipeline, you need a managed orchestrator to chain steps like model training, evaluation, and deployment, and Vertex AI Pipelines provides that with its Kubeflow Pipelines-based DAG execution.

Answer

Cloud Scheduler to trigger the pipeline on a schedule.

Answer

Cloud Functions to deploy the model.

Answer

Vertex AI Continuous Training service.

Question 15

You are tasked with building a robust ML pipeline that must be idempotent and handle data skew between training and serving. Which three practices should you implement?

Accepted Answer

Store intermediate data in Cloud Storage with unique run IDs.. Idempotent components ensure the same inputs produce the same outputs. Passing data via GCS URIs is a best practice. Skew detection should compare training data distribution with serving data. Using unique run IDs for outputs ensures idempotency. Avoiding in-memory data passing is important for large datasets.

Answer

Pass large datasets between components as serialized in-memory objects.

Answer

Use the same random seed for every run to ensure reproducibility.