AI-300 Microsoft Operationalizing Machine Learning and Generative AI Solutions Free Practice Exam Questions (2026 Updated)

No. Deleting an existing Azure Machine Learning kernel is neither required nor appropriate before adding another Jupyter kernel. Azure Machine Learning compute instances are explicitly designed to support multiple kernels simultaneously . Microsoft states that notebooks automatically discover the Jupyter kernels installed on the connected compute instance, allowing users to switch between them from the kernel selector.

To add a kernel, the supported process is to create a new Conda environment, activate it, install pip and ipykernel, and then register that environment by using python -m ipykernel install --user. Existing kernels do not need to be removed.

More importantly, Microsoft warns against deleting Conda environments or Jupyter kernels that you did not create , because doing so can damage Jupyter or JupyterLab functionality on the Azure Machine Learning compute instance. This makes removal of a built-in AzureML kernel particularly inappropriate as a prerequisite for creating another kernel.

The correct approach is therefore to leave the existing kernel intact and create a separate environment/kernel for the new workload.

Study Guide Reference: Design and implement an MLOps infrastructure — Azure Machine Learning notebook environments, kernel management, compute instances, and dependency isolation.

===============

Two distinct RAG problems require two distinct solutions. When retrieved results frequently include duplicated content from the same document, the cause is overlapping chunks receiving similarly high similarity scores. The solution is Maximum Marginal Relevance reranking, which diversifies the result set by penalizing results that are semantically similar to already-selected results, reducing redundancy. When retrieved chunks sometimes span unrelated policy sections, the cause is fixed-size character chunking that splits content without regard for semantic boundaries. The solution is semantic chunking — splitting on natural sentence or section boundaries — ensuring each chunk is semantically coherent and does not straddle unrelated content. Both problems must be addressed together: reranking alone solves duplication but not relevance; semantic chunking alone solves relevance but not duplication.

Microsoft Learn Reference Topic: Optimize RAG retrieval in Azure AI Search – Chunking strategies and MMR reranking

Microsoft ' s RAG optimization guidance identifies two high-impact, low-cost improvements for retrieval quality that do not require changing the embedding model. First, tuning chunk size and overlap (option A): chunk size determines how much context each retrieved piece contains — too large and irrelevant content dilutes the signal; too small and answers may be split across chunks. Adjusting these parameters requires only re-indexing the documents with zero additional compute cost. Second, implementing a re-ranker (option B): a re-ranker is a cross-encoder model that takes the top-N retrieved chunks and re-scores them based on their specific relevance to the query, significantly improving precision by filtering out contextually irrelevant chunks. Re-rankers add modest compute cost but are far cheaper than changing the embedding model, which would require re-embedding the entire knowledge base. Increasing token limits (option C) and optimizing embedding vector length (option D) do not address retrieval accuracy without an embedding model change.

Microsoft Learn Reference Topic: Optimize RAG pipelines – Chunk size tuning and re-ranking in Azure AI Search and Azure Machine Learning

Define how features are retrieved at inference: Feature retrieval specification

Ensure feature consistency between training and inference: Feature retrieval specification

Enable automated feature lookup in production: Feature retrieval specification

The feature retrieval specification is the key model-level contract for all three requirements. Microsoft defines it as a portable specification containing the exact list of features associated with a model, including the relevant feature store, feature set, and feature-set version . The same specification participates in both training and inference, making it the connective artifact across the model lifecycle.

During training, the feature retrieval specification is used to obtain the required feature values and generate training data. Microsoft requires the specification to be packaged with the model artifact when the model depends on feature-store features. At online inference time, the scoring script loads this packaged specification, resolves the feature list, and initializes online feature lookup before calling the feature store to retrieve current values.

This also prevents training-serving inconsistencies because the deployed model carries the feature dependencies and specific feature-set versions established during training, rather than relying on manually reconstructed production configuration. Microsoft explicitly states that packaging the retrieval specification with the model minimizes changes between training and inference workflows.

A feature set specification defines feature sources and transformations, a feature set asset provides managed versioning, and materialization precomputes feature values. None of those alone defines the model ' s complete retrieval contract at inference.

Study Guide Reference: Implement machine learning model lifecycle and operations — package a feature retrieval specification with the model artifact and operationalize feature-store-backed models.

A managed identity is the appropriate credential model when an Azure-hosted workload requires centrally controlled, non-user authentication to another service that supports Microsoft Entra identities. Managed identities eliminate developer-managed credentials because Azure manages the identity lifecycle and obtains short-lived Microsoft Entra tokens at runtime. Microsoft explicitly identifies managed identities as a mechanism for Azure-hosted applications to authenticate without storing credentials, and Azure DevOps supports managed identities for secure automation and repository-related access.

This satisfies the key MLOps security requirement: authentication is associated with the workload rather than individual developers. Administrators can grant the identity only the required repository permissions, enforcing least privilege while centralizing authorization. Microsoft also recommends Microsoft Entra-based authentication over higher-risk PAT-based approaches for Azure Repos.

Option A spreads a reusable private credential among developers and creates unnecessary exposure. Option B still introduces a PAT that must be stored, protected, rotated, and revoked. Option C depends on individual interactive authentication and is unsuitable for reliable automated MLOps workflows. Microsoft documents managed identities and workload identities as approaches that eliminate persistent secrets and reduce credential-management overhead.

Study Guide Reference: Design and implement an MLOps infrastructure — source-control integration, workload authentication, managed identities, least-privilege access, and credential management.

B and C provide the strongest architecture for reusable and predictable prompts. Separating persistent application instructions from dynamic user input establishes a clear instruction hierarchy . The stable portion of the prompt can define application behavior while channel-specific or request-specific content remains variable. This reduces coupling between user wording and the behavioral rules of the assistant and makes the prompt easier to reuse, version, test, and modify.

The system prompt should establish the model ' s role, behavioral boundaries, tone, communication style, response format, and other persistent rules. Microsoft specifically describes system messages as high-priority instructions used to define an assistant ' s role and boundaries and to set tone and communication style.

Fine-tuning for every conversational variation is unnecessarily expensive and contradicts the requirement to remain adaptable without retraining. Increasing temperature increases randomness and therefore works against consistent responses. Although repeating important instructions can sometimes mitigate recency effects, Microsoft treats this as an optional prompt-engineering technique to experiment with, not the primary architectural technique for producing reusable prompts.

Study Guide Reference: Design and implement a GenAIOps infrastructure — design and develop prompts, prompt management, prompt variants, system instructions, and reusable prompt structures.

===============

The two requirements target fundamentally different stages of the AI application lifecycle. Continuous runtime monitoring addresses production behavior — an application that intermittently times out needs real-time visibility into latency, error rates, and request volumes for every call in production. Tracing backed by Application Insights or OpenTelemetry exporters is correct here because it captures granular, per-request telemetry continuously without requiring manual triggering — it is always-on by design. Release-time validation addresses pre-release quality gates — before new prompts or model updates go live, you need to verify they have not introduced regressions. An evaluation pipeline in Microsoft Foundry ' s prompt flow is correct here: it runs the updated application against a predefined test dataset, scores outputs on quality and safety metrics, and passes or fails the release based on thresholds. Evaluation pipelines run on demand triggered by CI/CD, not continuously.

Microsoft Learn Reference Topic: Microsoft Foundry observability – Tracing for runtime monitoring vs. Evaluation pipelines for release gates

The Content Safety evaluator in Microsoft Foundry is specifically designed to detect hate speech, violence, sexual content, and self-harm content in model outputs, making it the most direct evaluator for the requirement of preventing hateful responses. It uses Azure AI Content Safety service under the hood, which is trained to classify content across these harmful categories with high accuracy. Protected Material (option A) evaluates whether outputs contain copyrighted or licensed material — not hate speech. Groundedness (option B) measures factual accuracy against a source context — completely irrelevant to hate detection. Indirect Attacks (option C) evaluates whether the model was manipulated via prompt injection to produce harmful content — this is a robustness metric, not a direct output quality measure. Content Safety must be configured first because detecting hate is the primary safety concern explicitly identified in the requirement.

Microsoft Learn Reference Topic: Content safety evaluators in Microsoft Foundry – Detect harmful content in AI model outputs

According to Microsoft ' s Azure Machine Learning SDK v2 documentation on command jobs, input arguments in the command string must exactly match the argument names defined in the Python script ' s argparse configuration. The script reads an argument named training_data with an underscore, but the proposed command passes --trainingdata without the underscore. Python ' s argparse treats these as completely different argument names, so the script would receive None for its --training_data argument, causing a runtime error. The placeholder syntax for referencing the input in the command string is correct, which is a point in this solution ' s favor. However, the argument name mismatch alone is sufficient to make this solution fail. The corrected command would use --training_data exactly matching the argument name defined in the script.

Microsoft Learn Reference Topic: Configure command job inputs and outputs – Azure Machine Learning SDK v2 reference

Data asset type: mltable

Scheme of the URL paths to storage locations: wasbs

The correct data asset type is mltable because the image collection is distributed across multiple storage locations . Microsoft identifies MLTable as the appropriate Azure Machine Learning data asset type when unstructured data such as images, text, or audio is spread across multiple storage locations. An MLTable definition can contain multiple file, folder, or pattern entries, allowing one logical data asset to reference image data residing in different storage accounts.

A uri_folder data asset is appropriate when the required unstructured files reside within a single folder path . It represents one folder location and therefore does not provide the same multi-path abstraction required by this scenario. uri_file represents only an individual file and is clearly unsuitable for a collection of images.

For the path scheme, Azure Machine Learning uses wasbs:// for Azure Blob Storage. Microsoft documents the canonical format as:

wasbs:// < container > @ < account > .blob.core.windows.net/ < path >

By comparison, abfss:// is used for Azure Data Lake Storage Gen2, while azureml:// identifies Azure Machine Learning datastore paths rather than direct Azure Blob Storage URLs.

Therefore, an MLTable containing multiple wasbs:// paths correctly represents image files distributed across the two Azure Blob Storage accounts.

Study Guide Reference: Design and implement an MLOps infrastructure — Azure Machine Learning data assets, MLTable, multi-location datasets, storage URI schemes, and Blob Storage integration.