PMLE Automating and Orchestrating ML Pipelines Practice Test 3 — 15 Questions

Question 1

A data science team wants to deploy a ML pipeline on Vertex AI Pipelines that includes a component to train a model using a custom container. The component should be reusable across different pipelines and accept hyperparameters as inputs. Which approach should they take?

Accepted Answer

Create a container component by specifying a container image and input/output artifacts using the Kubeflow Pipelines SDK.. Container components are designed for custom containers and can be defined once and reused across pipelines. Python function components are simpler but limited to Python code; they cannot easily encapsulate custom containers.

Answer

Package the training code as a Vertex AI Training custom job and call it from a Python function component.

Answer

Define a Python function component using @dsl.component and pass hyperparameters as function arguments.

Answer

Use a pre-built Google Cloud Pipeline Component for custom training and override the image.

Question 2

An ML engineer is building a pipeline on Vertex AI Pipelines and wants to pass a dataset artifact from one component to another without incurring additional cost for intermediate storage. How should they define the input and output types?

Accepted Answer

Use the Dataset artifact type available in the KFP SDK for inputs and outputs.. Option C is correct because using the KFP SDK's Dataset artifact type allows Vertex AI Pipelines to manage the data as a lineage-tracked artifact, automatically handling the underlying GCS storage reference without incurring additional intermediate storage costs. The artifact type enables the pipeline to pass the metadata (URI, type, etc.) between components efficiently, leveraging the native artifact management of the KFP SDK.

Answer

Store the data in BigQuery and pass the table reference.

Answer

Define output as a string and pass the GCS path manually.

Answer

Use in-memory Python objects as function return values.

Question 3

A team runs a Vertex AI pipeline daily. They notice that a component that downloads a file from a public URL always executes even when the URL and parameters haven't changed. They want to avoid unnecessary re-execution and reduce costs. What should they do?

Accepted Answer

Ensure the component is deterministic and that caching is not disabled for that component.. Vertex AI Pipelines caching relies on a cache key based on component inputs, source code, and base image digest. If a component always re-executes, it may be because caching is explicitly disabled for that component (via 'disable_caching=True') or the component is non-deterministic (e.g., generating random values or including timestamps). Option A addresses both: ensure the component is deterministic and verify caching is not disabled. Option B is wrong because 'enable_caching=True' on the pipeline decorator is not the correct API; caching is enabled by default and controlled per task via 'dsl.CachingOptions' or the 'disable_caching' parameter. Option C is irrelevant—using a pre-built component does not bypass caching issues if caching is disabled or the component is non-deterministic. Option D is unnecessary; custom cache keys are not required if the default caching works; the problem is either caching disabled or non-determinism.

Answer

Enable pipeline caching by setting 'enable_caching=True' on the pipeline decorator.

Answer

Use a pre-built Google Cloud Pipeline Component for file download.

Answer

Use the 'dsl.CachingOptions' to set a custom cache key for the component.

Question 4

An organization wants to trigger a Vertex AI pipeline whenever new data arrives in a Cloud Storage bucket. Which approach should they use?

Accepted Answer

Use Cloud Functions triggered by Cloud Storage events to call the Vertex AI pipeline API.. Option D is correct because Cloud Functions can be directly triggered by Cloud Storage events (e.g., `google.storage.object.finalize`) and can then call the Vertex AI pipeline API using the Cloud SDK or client libraries. This provides a serverless, event-driven architecture that reacts immediately to new data without polling or additional infrastructure.

Answer

Configure a Vertex AI pipeline trigger directly on the bucket using the GCP Console.

Answer

Use Pub/Sub notifications from the bucket and a Dataflow job to start the pipeline.

Answer

Set up a Cloud Scheduler job that runs every minute and checks for new files in the bucket.

Question 5

An ML engineer is using Cloud Composer (Airflow) to orchestrate a ML workflow. They need to run a Vertex AI pipeline as one of the tasks in the DAG. Which Airflow operator should they use?

Accepted Answer

VertexAIPipelineRunOperator or the Google Cloud Pipeline operator.. Option B is correct because Cloud Composer (Airflow) natively supports the `VertexAIPipelineRunOperator` (or its alias `GoogleCloudPipelineOperator`), which is specifically designed to trigger and monitor a Vertex AI pipeline run as a task within a DAG. This operator handles authentication, pipeline job submission, and status polling without requiring custom code, making it the idiomatic choice for orchestrating Vertex AI pipelines from Airflow.

Answer

BigQueryOperator to run the pipeline as a query.

Answer

PythonOperator with a custom script using the google-cloud-aiplatform library.

Answer

VertexAIPipelineRunOperator (or Airflow's GCSToGCSOperator) for pipeline orchestration.

Question 6

A team has a pipeline that trains a model and then evaluates it. They want to conditionally deploy the model to a staging endpoint only if evaluation metrics exceed a threshold. Which KFP feature should they use?

Accepted Answer

Use dsl.Condition (deprecated) or dsl.If to check metrics and conditionally run deployment.. Option A is correct because KFP provides `dsl.Condition` (deprecated) and `dsl.If` as first-class pipeline constructs to conditionally execute pipeline components based on runtime metrics or other pipeline outputs. By wrapping the deployment step inside a `dsl.If` block that checks whether evaluation metrics exceed a threshold, the pipeline can deploy the model to a staging endpoint only when the condition is met, avoiding unnecessary deployments for underperforming models.

Answer

Use dsl.ParallelFor to evaluate and deploy in parallel.

Answer

Use an exit handler to deploy regardless of metrics.

Answer

Split the pipeline into two separate pipelines and run the second only if metrics are good.

Question 7

What is the primary benefit of using pipeline caching in Vertex AI Pipelines?

Accepted Answer

It reduces execution time and cost by reusing unchanged component outputs.. Pipeline caching in Vertex AI Pipelines automatically detects when a component's inputs and code have not changed from a previous execution and reuses the cached output artifacts. This avoids redundant computation, directly reducing both execution time and cost by skipping re-execution of unchanged steps.

Answer

It encrypts data at rest.

Answer

It automatically scales the pipeline resources.

Answer

It enables parallel execution of components.

Question 8

An engineer needs to compile a Kubeflow Pipeline defined in Python to a JSON format that can be run on Vertex AI Pipelines. Which command should they use?

Accepted Answer

kfp.compiler.Compiler().compile(pipeline_func, 'pipeline.json'). Option A is correct because the Kubeflow Pipelines SDK provides the `kfp.compiler.Compiler().compile()` method to convert a Python-based pipeline function into a JSON or YAML format that is compatible with Vertex AI Pipelines. This JSON representation defines the pipeline's components, dependencies, and execution graph, enabling it to be submitted to Vertex AI for orchestration. The `compile()` method is the standard way to produce a portable pipeline specification from Python code.

Answer

gcloud ai pipelines compile command.

Answer

kfp.Client().upload_pipeline()

Answer

dsl.pipeline decorator automatically compiles at runtime.

Question 9

A company has a CI/CD pipeline that retrains a model every time new training data is available. They want to automatically deploy the new model to production only if it passes a set of evaluation tests on a staging environment. Which approach best implements this?

Accepted Answer

Implement a two-stage pipeline: train and deploy to staging, run evaluation tests, and if passed, deploy to production using conditional logic.. A continuous delivery pipeline with evaluation gates uses separate stages: staging deployment → evaluation → promotion to production.

Answer

Use Cloud Build to trigger a training job and then a separate deployment job without evaluation.

Answer

Use a single Vertex AI pipeline that trains and deploys to staging, then manually promote.

Answer

Train and deploy directly to production in one pipeline.

Question 10

Which of the following is a best practice when designing idempotent pipeline components in Vertex AI?

Accepted Answer

Pass data through Cloud Storage URIs rather than in-memory.. Passing data through Cloud Storage URIs ensures that component outputs are stored persistently and can be retrieved by downstream components, even if the original component instance is terminated or scaled down. This aligns with the principle of idempotency because the same input will always produce the same output stored at the same URI, and re-running the component will not cause side effects or data loss. In contrast, in-memory data is ephemeral and tied to a specific runtime instance, breaking idempotency across retries or parallel executions.

Answer

Use global variables to share state between components.

Answer

Write component outputs to a database with timestamps.

Answer

Use the same output name for all runs to avoid duplication.

Question 11

An ML team wants to run a hyperparameter tuning job on Vertex AI using a pre-built pipeline component. Which component should they use?

Accepted Answer

HyperparameterTuningJobRunOp. The HyperparameterTuningJobRunOp is the correct pre-built Vertex AI pipeline component specifically designed to launch a hyperparameter tuning job. It wraps the Vertex AI HyperparameterTuningJob API, allowing you to specify the worker pool spec, metric target, and parameter specifications directly within a Kubeflow Pipelines (KFP) or Vertex AI Pipelines orchestration context.

Answer

AutoMLTabularTrainingJobRunOp

Answer

CustomTrainingJobRunOp with hyperparameter arguments.

Answer

ModelTrainComponent

Question 12

What is the purpose of the 'importer' component in Vertex AI Pipelines?

Accepted Answer

To import existing ML artifacts (e.g., models, datasets) into a pipeline as inputs.. The importer component allows you to bring existing artifacts (e.g., a model in Vertex AI Model Registry) into a pipeline for downstream use.

Answer

To import data from external sources into the pipeline.

Answer

To import Python libraries into the pipeline environment.

Answer

To import pipeline definitions from other projects.

Question 13

Your organization wants to automate the retraining of a model when new data is available and also on a weekly schedule. Which TWO services would you use together to achieve this? (Choose two.)

Accepted Answer

Cloud Functions. Cloud Scheduler (E) is used to trigger the retraining on a weekly schedule by sending a message to a Pub/Sub topic or making an HTTP request. Cloud Functions (A) is the serverless compute service that executes the retraining code in response to that trigger, and it can also be triggered directly when new data arrives (e.g., via Cloud Storage or Pub/Sub). Together, they provide both event-driven and scheduled automation without managing infrastructure.

Answer

Cloud Composer

Answer

Cloud Tasks

Answer

Dataflow

Question 14

A team is designing a ML pipeline that includes training, evaluation, and conditional deployment. They want to use Vertex AI Pipelines. Which THREE concepts should they use? (Choose three.)

Accepted Answer

Artifact types (e.g., Model, Metrics) for passing outputs. dsl.If for conditionals, Artifacts for passing model/data, and pre-built Vertex AI components for training.

Answer

Manual approval via Cloud Console

Answer

Cloud SQL for storing intermediate results

Question 15

An ML engineer is building a continuous training pipeline that retrains a model when new data arrives. The pipeline should also detect skew between training and serving data. Which TWO Google Cloud services should they use? (Choose two.)

Accepted Answer

Vertex AI Model Monitoring. Vertex AI Model Monitoring (B) is correct because it is purpose-built to detect skew between training and serving data by continuously comparing feature distributions and alerting on statistically significant drift. Vertex AI Pipelines (D) is correct because it provides a serverless, scalable orchestration service for building continuous training pipelines that automatically retrain models when new data arrives, integrating with Cloud Build and other services.

Answer

Cloud Logging

Answer

Cloud Functions

Answer

Cloud Monitoring