AI-900 Practice Question: Describe fundamental principles of machine learning on Azure
What is 'model versioning' and why is it essential in MLOps?
⚠ Common exam trap
It's easy for candidates to confuse model versioning with data versioning or environment versioning, but the question specifically asks about tracking the trained model artifact itself for rollback, A/B testing, auditing, and reproducibility.
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
✓
Tracking each iteration of a trained model for rollback, A/B testing, auditing, and reproducibility
Model versioning is the practice of tracking each iteration of a trained model, including its hyperparameters, training data snapshot, and evaluation metrics. In MLOps, it is essential because it enables rollback to a previous model if a new version performs poorly, supports A/B testing by comparing multiple model versions in production, provides an audit trail for compliance, and ensures reproducibility by capturing the exact code, data, and environment used to train each version.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Updating the Python version used to run ML training scripts
Why it's wrong here
Updating the Python version used to run ML training scripts is environment configuration, not model versioning: it manages the runtime dependencies and compute environment for a training run. While the Python interpreter version can affect training results, updating it does not register or track the trained model artefacts themselves. In Azure ML, environment changes are captured in environment objects and run configurations, but model versioning specifically records each iteration of the model in the model registry for rollback, audit, and reproduction.
- ✓
Tracking each iteration of a trained model for rollback, A/B testing, auditing, and reproducibility
Why this is correct
Trackting each iteration of a trained model is the core definition of model versioning: each training run can produce a distinct model artefact registered with a unique version, along with metadata such as hyperparameters, metrics, and source code commit. This enables safe rollback to a prior model when a new version underperforms, controlled A/B testing between two candidate models, a full audit trail for regulatory compliance, and exact reproducibility of any deployed model. In Azure ML, the model registry manages these versions, supports lifecycle states, and allows promotion of specific iterations to staging or production.
- ✗
Releasing new features of the Azure ML service as versioned API updates
Why it's wrong here
Releasing new features of the Azure ML service as versioned API updates is product-level versioning performed by Microsoft as the platform evolves; it is not your practice of tracking your own trained model iterations. Even if you consume a newer service API version, that does not create a versioned record of your model's parameters, training data, or performance metrics. Model versioning is an MLOps discipline applied to artefacts you produce, such as serialized models and their lineage, which is independent of the service's own release cycle.
- ✗
Managing multiple versions of training data used by different model experiments
Why it's wrong here
Managing multiple versions of training data is a related but separate practice called data versioning: it tracks datasets and their lineage, not the trained model artefacts themselves. A data version identifies a specific snapshot of inputs, whereas a model version identifies the fitted parameters, weights, and metadata produced from a training run. In Azure ML, datasets are registered and versioned independently from models in the model registry, so dataset versioning cannot by itself provide rollback, A/B testing, or audit trails for model iterations.
Go deeper
Related to this question
Learn chapter
Features, Labels, and Training Data
Key term
Training data
Training data is a set of examples used to teach a machine learning model how to make predictions or decisions.
Key term
Model
In IT and AI, a model is a trained mathematical representation that learns patterns from data to make predictions or decisions.
About these practice questions
Courseiva writes every AI-900 question from scratch — 985 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-900 practice question is part of Courseiva's free Microsoft 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 AI-900 exam.