Practice Exams:

Microsoft AI-300 Dumps

EXAM: Operationalizing Machine Learning and Generative AI Solutions
Verified By Experts
Premium Bundle
$49.99

AI-300 Premium ETE File


  • Premium File 213 Questions & Answers
Whats Included:
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

Microsoft AI-300 Practice Test 9

Question 80

A research consortium wants approved model versions reused across staging and production Azure ML workspaces while keeping their compute, storage and access independently governed. Which combined implementation and verification plan is the best match for this workload?

  1. Implement: Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces. Verify: Register one disposable versioned component, consume it from a separate authorized workspace and confirm a denied identity cannot resolve restricted registry content.
  2. Implement: Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces. Verify: Run the same training component in a fresh compute cluster using the pinned environment, inspect package and image versions, and compare the expected deterministic validation metrics.
  3. Implement: Publish a versioned Azure Machine Learning environment with reviewed Python, framework and system dependencies, and reference that immutable environment version from submitted jobs. Verify: Register one disposable versioned component, consume it from a separate authorized workspace and confirm a denied identity cannot resolve restricted registry content.
  4. Implement: Package a training step as a versioned Azure Machine Learning command component with declared input and output types, environment and command rather than embedding hidden file-system assumptions. Verify: Invoke the same registered component from two test pipelines with different versioned input assets and verify each run emits a correctly typed output artifact in its isolated location.

Answer & Explanations

Correct answer: A

A.

The question is set in this situation: A research consortium wants approved model versions reused across staging and production Azure ML workspaces while keeping their compute, storage and access independently governed. Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces. Confirm it through this test: Register one disposable versioned component, consume it from a separate authorized workspace and confirm a denied identity cannot resolve restricted registry content. Registries can support reuse across workspace boundaries without transferring ownership of every source workspace, but access and asset version resolution still require supported identity and regional configuration.

B.

The question is set in this situation: A research consortium wants approved model versions reused across staging and production Azure ML workspaces while keeping their compute, storage and access independently governed. The action is applicable: Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces. But its proposed test—Run the same training component in a fresh compute cluster using the pinned environment, inspect package and image versions, and compare the expected deterministic validation metrics.—addresses “Pin Azure Machine Learning environments for repeatable training dependencies”. This requirement is confirmed instead by: Register one disposable versioned component, consume it from a separate authorized workspace and confirm a denied identity cannot resolve restricted registry content.

C.

The question is set in this situation: A research consortium wants approved model versions reused across staging and production Azure ML workspaces while keeping their compute, storage and access independently governed. The test would be relevant: Register one disposable versioned component, consume it from a separate authorized workspace and confirm a denied identity cannot resolve restricted registry content. But the offered implementation—Publish a versioned Azure Machine Learning environment with reviewed Python, framework and system dependencies, and reference that immutable environment version from submitted jobs.—addresses “Pin Azure Machine Learning environments for repeatable training dependencies”. The action needed here is: Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces.

D.

The question is set in this situation: A research consortium wants approved model versions reused across staging and production Azure ML workspaces while keeping their compute, storage and access independently governed. This action/test pair addresses “Create reusable Azure Machine Learning components with typed inputs and outputs”. Package a training step as a versioned Azure Machine Learning command component with declared input and output types, environment and command rather than embedding hidden file-system assumptions. Invoke the same registered component from two test pipelines with different versioned input assets and verify each run emits a correctly typed output artifact in its isolated location. The case instead calls for: Publish approved models, components or environments into a shared Azure Machine Learning registry with explicit version identifiers, reader permissions and region requirements for consuming workspaces.

Question 81

A regulated insurer needs identical nonproduction and production Azure ML foundations with different private network addresses and documented approvals. Which Azure AI operations action best addresses this requirement?

  1. Provision the Azure Machine Learning workspace in the approved region with its required storage, Key Vault, monitoring dependencies and project-specific resource ownership, then verify workspace provisioning completes.
  2. Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access.
  3. Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.
  4. Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments.

Answer & Explanations

Correct answer: C

A.

Provision the Azure Machine Learning workspace in the approved region with its required storage, Key Vault, monitoring dependencies and project-specific resource ownership, then verify workspace provisioning completes. This addresses “Create an Azure Machine Learning workspace with governed dependent resources”, where the governing principle is: The workspace is the management boundary for jobs and registered ML assets while supporting Azure services hold artifacts, credentials and telemetry; a bare resource-group label does not create working ML dependencies. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

B.

Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access. This addresses “Validate private network and DNS dependencies of restricted Azure ML workspaces”, where the governing principle is: Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

C.

Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies. Verify the result with this test: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

D.

Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments. This addresses “Use GitHub Actions workload identity federation for secure ML provisioning”, where the governing principle is: OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Question 82

A regulated insurer needs identical nonproduction and production Azure ML foundations with different private network addresses and documented approvals. The engineer proposes the following change: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Which mechanism explains why it works?

  1. Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage.
  2. The workspace is the management boundary for jobs and registered ML assets while supporting Azure services hold artifacts, credentials and telemetry; a bare resource-group label does not create working ML dependencies.
  3. OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment.
  4. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

Answer & Explanations

Correct answer: D

A.

Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage. That explains “Validate private network and DNS dependencies of restricted Azure ML workspaces”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

B.

The workspace is the management boundary for jobs and registered ML assets while supporting Azure services hold artifacts, credentials and telemetry; a bare resource-group label does not create working ML dependencies. That explains “Create an Azure Machine Learning workspace with governed dependent resources”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

C.

OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment. That explains “Use GitHub Actions workload identity federation for secure ML provisioning”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

D.

Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies. The corresponding implementation is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Question 83

A regulated insurer needs identical nonproduction and production Azure ML foundations with different private network addresses and documented approvals. After implementing this change—Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.—which test most directly validates its intended outcome?

  1. From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction.
  2. Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  3. Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege.
  4. Create a small experiment in the new workspace, write a disposable data artifact to approved storage and confirm logs and controlled secret references resolve for a nonadministrator test identity.

Answer & Explanations

Correct answer: B

A.

From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction. That would help test “Validate private network and DNS dependencies of restricted Azure ML workspaces”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

B.

Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. This check is decisive because infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

C.

Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege. That would help test “Use GitHub Actions workload identity federation for secure ML provisioning”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

D.

Create a small experiment in the new workspace, write a disposable data artifact to approved storage and confirm logs and controlled secret references resolve for a nonadministrator test identity. That would help test “Create an Azure Machine Learning workspace with governed dependent resources”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

Question 84

A regulated insurer needs identical nonproduction and production Azure ML foundations with different private network addresses and documented approvals. Which proposed shortcut most directly risks loss of registered models and unnecessary outages during repeated infrastructure updates and should therefore be rejected?

  1. Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.
  2. Disable public access to every Azure ML dependency before adding private DNS and network routes, then assume jobs will still discover all endpoints automatically.
  3. Create only an empty Azure resource group named MachineLearning and assume the Machine Learning workspace, storage, monitoring and vault services appear implicitly.
  4. Commit a service principal client secret into a public workflow YAML file and claim repository branch protection makes leaked permanent credentials safe.

Answer & Explanations

Correct answer: A

A.

Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short. The foreseeable failure is loss of registered models and unnecessary outages during repeated infrastructure updates. The safer response is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

B.

Disable public access to every Azure ML dependency before adding private DNS and network routes, then assume jobs will still discover all endpoints automatically. That shortcut risks training jobs failing to fetch data or persist outputs after incomplete private networking. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

C.

Create only an empty Azure resource group named MachineLearning and assume the Machine Learning workspace, storage, monitoring and vault services appear implicitly. That shortcut risks unusable workspace operations after a deployment that never provisioned required machine learning resources. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

D.

Commit a service principal client secret into a public workflow YAML file and claim repository branch protection makes leaked permanent credentials safe. That shortcut risks unauthorized machine learning infrastructure deployments after long-lived pipeline credentials are exposed. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

Question 85

A regulated insurer needs identical nonproduction and production Azure ML foundations with different private network addresses and documented approvals. Which combined implementation and verification plan is the best match for this workload?

  1. Implement: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Verify: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  2. Implement: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Verify: Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege.
  3. Implement: Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments. Verify: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  4. Implement: Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access. Verify: From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction.

Answer & Explanations

Correct answer: A

A.

Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Confirm it through this test: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

B.

The action is applicable: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. But its proposed test—Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege.—addresses “Use GitHub Actions workload identity federation for secure ML provisioning”. This requirement is confirmed instead by: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

C.

The test would be relevant: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. But the offered implementation—Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments.—addresses “Use GitHub Actions workload identity federation for secure ML provisioning”. The action needed here is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

D.

This action/test pair addresses “Validate private network and DNS dependencies of restricted Azure ML workspaces”. Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access. From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction. The case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Question 86

A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Which Azure AI operations action best addresses this requirement?

  1. Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments.
  2. Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.
  3. Publish a versioned Azure Machine Learning environment with reviewed Python, framework and system dependencies, and reference that immutable environment version from submitted jobs.
  4. Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access.

Answer & Explanations

Correct answer: B

A.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments. This addresses “Use GitHub Actions workload identity federation for secure ML provisioning”, where the governing principle is: OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

B.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies. Verify the result with this test: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

C.

Publish a versioned Azure Machine Learning environment with reviewed Python, framework and system dependencies, and reference that immutable environment version from submitted jobs. This addresses “Pin Azure Machine Learning environments for repeatable training dependencies”, where the governing principle is: A versioned environment defines the container and package runtime needed for training, whereas installing libraries ad hoc on a notebook does not recreate those dependencies inside managed job compute. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

D.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access. This addresses “Validate private network and DNS dependencies of restricted Azure ML workspaces”, where the governing principle is: Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage. The stated case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Question 87

A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. The engineer proposes the following change: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Which mechanism explains why it works?

  1. A versioned environment defines the container and package runtime needed for training, whereas installing libraries ad hoc on a notebook does not recreate those dependencies inside managed job compute.
  2. OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment.
  3. Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage.
  4. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

Answer & Explanations

Correct answer: D

A.

A versioned environment defines the container and package runtime needed for training, whereas installing libraries ad hoc on a notebook does not recreate those dependencies inside managed job compute. That explains “Pin Azure Machine Learning environments for repeatable training dependencies”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

B.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. OpenID Connect federation supplies short-lived tokens conditioned on workflow claims, avoiding a stored long-lived Azure client secret in repository settings and requiring an appropriate Azure role assignment. That explains “Use GitHub Actions workload identity federation for secure ML provisioning”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

C.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Creating a workspace private endpoint alone does not guarantee the compute nodes can resolve or reach every linked artifact service needed for experiment submission and output storage. That explains “Validate private network and DNS dependencies of restricted Azure ML workspaces”, not the mechanism behind the described change. The relevant principle is: Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

D.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies. The corresponding implementation is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Question 88

A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. After implementing this change—Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.—which test most directly validates its intended outcome?

  1. From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction.
  2. Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege.
  3. Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  4. Run the same training component in a fresh compute cluster using the pinned environment, inspect package and image versions, and compare the expected deterministic validation metrics.

Answer & Explanations

Correct answer: C

A.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction. That would help test “Validate private network and DNS dependencies of restricted Azure ML workspaces”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

B.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege. That would help test “Use GitHub Actions workload identity federation for secure ML provisioning”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

C.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. This check is decisive because infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

D.

Run the same training component in a fresh compute cluster using the pinned environment, inspect package and image versions, and compare the expected deterministic validation metrics. That would help test “Pin Azure Machine Learning environments for repeatable training dependencies”, not the outcome required in this scenario. The relevant acceptance check is: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

Question 89

A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Which proposed shortcut most directly risks loss of registered models and unnecessary outages during repeated infrastructure updates and should therefore be rejected?

  1. Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.
  2. Commit a service principal client secret into a public workflow YAML file and claim repository branch protection makes leaked permanent credentials safe.
  3. Disable public access to every Azure ML dependency before adding private DNS and network routes, then assume jobs will still discover all endpoints automatically.
  4. Install arbitrary latest packages into a personal notebook and assume every managed training job will inherit those notebook-side packages automatically.

Answer & Explanations

Correct answer: A

A.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short. The foreseeable failure is loss of registered models and unnecessary outages during repeated infrastructure updates. The safer response is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

B.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Commit a service principal client secret into a public workflow YAML file and claim repository branch protection makes leaked permanent credentials safe. That shortcut risks unauthorized machine learning infrastructure deployments after long-lived pipeline credentials are exposed. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

C.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Disable public access to every Azure ML dependency before adding private DNS and network routes, then assume jobs will still discover all endpoints automatically. That shortcut risks training jobs failing to fetch data or persist outputs after incomplete private networking. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

D.

Install arbitrary latest packages into a personal notebook and assume every managed training job will inherit those notebook-side packages automatically. That shortcut risks training job failures from environment drift between experimentation and managed compute. Here the specific shortcut is: Run an unguarded script that destroys and recreates the workspace and training artifact storage on every deployment to keep the code short.

Question 90

A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Which combined implementation and verification plan is the best match for this workload?

  1. Implement: Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access. Verify: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  2. Implement: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Verify: From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction.
  3. Implement: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Verify: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.
  4. Implement: Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments. Verify: Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege.

Answer & Explanations

Correct answer: C

A.

The test would be relevant: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. But the offered implementation—Configure supported Azure Machine Learning private endpoint and network access settings along with dependent storage and vault connectivity, private DNS and outbound egress requirements before disabling public access.—addresses “Validate private network and DNS dependencies of restricted Azure ML workspaces”. The action needed here is: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

B.

The action is applicable: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. But its proposed test—From an approved compute subnet run a small training job that reads its private datastore, resolves required service names and writes logs while public network access follows the approved restriction.—addresses “Validate private network and DNS dependencies of restricted Azure ML workspaces”. This requirement is confirmed instead by: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation.

C.

The question is set in this situation: A university wants the research infrastructure team to reproduce a Machine Learning workspace after a region migration without relying on screenshots of past portal settings. Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers. Confirm it through this test: Deploy the template to a disposable resource group twice with unchanged parameters and verify the second execution introduces no unexpected resource replacement or network relaxation. Infrastructure-as-code reconciles desired resource declarations across deployments, while copying portal clicks by hand leaves undocumented differences in networking, identity and workspace dependencies.

D.

This action/test pair addresses “Use GitHub Actions workload identity federation for secure ML provisioning”. Configure a trusted GitHub Actions workflow to authenticate to Azure through a narrowly scoped federated identity credential and deploy reviewed ML resources only from authorized repositories or environments. Trigger the permitted deployment workflow in a sandbox, confirm an authorized branch can obtain a token and make intended resource changes, and verify an untrusted pull request lacks equivalent privilege. The case instead calls for: Define the approved Machine Learning workspace, network policies and dependent services in parameterized Bicep templates and deploy with Azure CLI using environment-specific configuration and repeatable resource identifiers.

Operationalizing Machine Learning and Generative AI Solutions (AI-300) Practice Tests