easyMultiple Choice
PMLE Practice Question: An ML engineer is designing a CI/CD pipeline for…
An ML engineer is designing a CI/CD pipeline for ML models using Cloud Build and Cloud Deploy. They want to automatically test model performance on a validation set before promoting to production. Which step should be included in the CI/CD pipeline?
⚠ Common exam trap
Google Cloud often tests the distinction between code testing (unit tests) and model validation (performance metrics), leading candidates to choose A because they conflate software testing with ML evaluation.
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
✓
Run a Vertex AI Pipeline for model evaluation and register the model only if metrics exceed thresholds
It directly integrates model evaluation into the CI/CD pipeline using Vertex AI Pipelines, which allows automated validation of model performance against predefined thresholds before promotion. This ensures that only models meeting quality criteria are deployed, aligning with MLOps best practices for gated promotions.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Run unit tests on the training code
Why it's wrong here
Unit tests on training code verify code correctness, not model quality on held-out data, so they cannot gate promotion on validation performance. It is tempting because unit tests belong in every CI pipeline, and would be correct for catching code regressions before training runs.
- ✗
Use Cloud Composer to schedule evaluation
Why it's wrong here
Cloud Composer schedules workflows; it does not itself evaluate a model against a validation set within the Cloud Build and Cloud Deploy pipeline. It is tempting because Composer orchestrates ML pipelines, and would be correct for recurring retraining or batch scoring jobs rather than pre-promotion validation.
- ✗
Deploy to production immediately after training
Why it's wrong here
Deploying straight after training skips the validation gate entirely, so no performance check occurs before production. This suits rapid prototyping where rollback is trivial, but the stem explicitly requires automated validation-set testing prior to promotion.
- ✗
Train the model in the CI/CD pipeline
Why it's wrong here
Training inside the pipeline consumes compute without validating the artefact the stem requires; the pipeline must evaluate an already-trained model on a held-out validation set before promotion. Training belongs in a separate upstream job or retraining workflow, not the promotion gate.
- ✓
Run a Vertex AI Pipeline for model evaluation and register the model only if metrics exceed thresholds
Why this is correct
Embedding a Vertex AI Pipeline evaluation step gates promotion on measured validation metrics, so Cloud Deploy only releases a model whose performance exceeds defined thresholds. This satisfies the stem's requirement to test performance automatically before production, rather than relying on manual review or post-deployment monitoring.
Go deeper
Related to this question
About these practice questions
One of 775 original PMLE practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This PMLE practice question is part of Courseiva's free Google Cloud certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the PMLE exam.