McAfee Secure

Exam Code: AI-300

Exam Name: Operationalizing Machine Learning and Generative AI Solutions

Certification Provider: Microsoft

Corresponding Certification: Microsoft Certified: Machine Learning Operations Engineer Associate

Microsoft AI-300 Questions & Answers

Study with Up-To-Date REAL Exam Questions and Answers from the ACTUAL Test

213 Questions & Answers with Testing Engine

"Operationalizing Machine Learning and Generative AI Solutions Exam", also known as AI-300 exam, is a Microsoft certification exam.

Pass your tests with the always up-to-date AI-300 Exam Engine. Your AI-300 training materials keep you at the head of the pack!

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Was: $137.49
Now: $124.99

Microsoft AI-300 Practice Test 4

Question 30

A shared job queue waits although an AML compute cluster has a maximum node count of one. The team is preparing a controlled change and wants evidence that it corrects the root cause. The team needs evidence that the cluster allocates additional eligible workers for parallel jobs. What is the best next action for this Azure AI workload?

  1. Create an AML compute cluster with appropriate min and max node settings
  2. Point the job at a compatible GPU-backed compute target with available quota
  3. Increase the training cluster's supported maximum nodes after checking quota and policy
  4. Request eligible quota or select an available supported region and VM family before resubmitting

Explanation

Correct Answer

C:

A maximum node limit bounds parallel worker capacity regardless of the workload queue size. On retest, the cluster allocates additional eligible workers for parallel jobs.

Incorrect Answers

A:

An AML compute cluster supports job-oriented scalable worker capacity rather than a fixed personal notebook VM.

B:

The selected Azure ML compute target determines accessible accelerator hardware.

D:

A quota or regional capacity problem is not corrected by increasing script retries. A maximum node limit bounds parallel worker capacity regardless of the workload queue size.

Question 31

An administrator tries to resize the VM family of an existing compute instance in place. Before rollout, the engineer must confirm that the replacement reports the target VM size without changing the old instance identity. What is the best next action for this Azure AI workload?

  1. Repair the approved outbound service connectivity and private network dependency paths
  2. Create or assign an appropriate individual compute instance instead of sharing the existing user's session
  3. Use the documented private network connection method and supported SSH configuration rather than exposing an unrestricted public port
  4. Create a replacement compute instance with the required supported VM size and migrate approved user work

Explanation

Correct Answer

D:

The compute instance virtual machine size cannot be changed after instance creation. After the change, the replacement reports the target VM size without changing the old instance identity.

Incorrect Answers

A:

Private compute still needs documented access to required package and AML services.

B:

An AML compute instance interactive development surface is assigned to one user.

C:

Restricted compute access and network boundary policy determine remote connectivity options. The compute instance virtual machine size cannot be changed after instance creation.

Question 32

A compute instance boots in a private subnet but cannot pull required packages from its approved services. The next test must distinguish a real fix from a configuration change that leaves the symptom intact. After remediation, the operator will verify that the provisioning test installs dependencies through the permitted path. What is the most appropriate corrective action?

  1. Create a replacement compute instance with the required supported VM size and migrate approved user work
  2. Use the documented private network connection method and supported SSH configuration rather than exposing an unrestricted public port
  3. Enable idle shutdown or a suitable automatic stop schedule on the compute instance
  4. Repair the approved outbound service connectivity and private network dependency paths

Explanation

Correct Answer

D:

Private compute still needs documented access to required package and AML services. After the change, the provisioning test installs dependencies through the permitted path.

Incorrect Answers

A:

The compute instance virtual machine size cannot be changed after instance creation.

B:

Restricted compute access and network boundary policy determine remote connectivity options.

C:

Compute instances incur running compute charges unless stopped or scheduled.

Question 33

A cluster job starts repeatedly on CPU hosts despite requiring CUDA acceleration. The project owner needs a repeatable outcome rather than another temporary workaround. The next run should prove that job telemetry reports a supported GPU device for the intended workload. Which decision addresses the issue with the fewest unrelated changes?

  1. Increase the training cluster's supported maximum nodes after checking quota and policy
  2. Plan quota using the instance's reserved restart capacity and separately monitor regional limits
  3. Point the job at a compatible GPU-backed compute target with available quota
  4. Request additional quota for the existing CPU-only compute target without selecting a GPU-backed VM family.

Explanation

Correct Answer

C:

The selected Azure ML compute target determines accessible accelerator hardware. The observable result is that job telemetry reports a supported GPU device for the intended workload.

Incorrect Answers

A:

More workers of the current CPU-only VM type still cannot execute CUDA kernels. The job needs an accelerator-backed target.

B:

Reserved core quota on a stopped development instance is separate from the hardware used by this training job.

D:

Additional quota for the existing CPU-only compute target does not supply GPU hardware to the running training job.

Question 34

A research team stores large temporary scratch datasets on the operating-system disk until it fills. The team can modify the relevant resource but must not broaden unrelated permissions or access. The team needs evidence that a full dry run stays within the compute disk and data-store limits. Which change most directly addresses the observed behavior?

  1. Create or assign an appropriate individual compute instance instead of sharing the existing user's session
  2. Configure the job to use appropriate scratch or mounted data storage for its expected size
  3. Use the documented private network connection method and supported SSH configuration rather than exposing an unrestricted public port
  4. Create a replacement compute instance with the required supported VM size and migrate approved user work

Explanation

Correct Answer

B:

OS disk capacity is not an unlimited workspace for large intermediate training artifacts. After the change, a full dry run stays within the compute disk and data-store limits.

Incorrect Answers

A:

Giving the researcher a separate development instance does not move large intermediate data off the disk that is filling.

C:

Changing SSH reachability affects remote management, not where temporary training data consumes disk space.

D:

A replacement VM size is a larger change and does not by itself place scratch data on an appropriately sized or mounted volume.

Question 35

A private compute instance lacks the SSH capability expected by a remote debugging procedure. The team wants to retain the approved design while correcting this specific failure. The team needs evidence that an authorized operator connects from the approved path with no public SSH exposure. Which decision addresses the issue with the fewest unrelated changes?

  1. Create an Azure Machine Learning compute instance assigned to that researcher
  2. Repair the approved outbound service connectivity and private network dependency paths
  3. Use the documented private network connection method and supported SSH configuration rather than exposing an unrestricted public port
  4. Enable idle shutdown or a suitable automatic stop schedule on the compute instance

Explanation

Correct Answer

C:

Restricted compute access and network boundary policy determine remote connectivity options. Check that an authorized operator connects from the approved path with no public SSH exposure.

Incorrect Answers

A:

A compute instance is a managed single-user development environment unlike a shared training cluster.

B:

Private compute still needs documented access to required package and AML services.

D:

Compute instances incur running compute charges unless stopped or scheduled.

Question 36

A data scientist attempts to open another user's AML compute instance interactive notebook. The operations team wants to preserve traceability while changing the minimum necessary configuration. A release test must show that both users work on separate authorized instances without sharing privileged notebooks. What is the most appropriate corrective action?

  1. Create a replacement compute instance with the required supported VM size and migrate approved user work
  2. Use the documented private network connection method and supported SSH configuration rather than exposing an unrestricted public port
  3. Enable idle shutdown or a suitable automatic stop schedule on the compute instance
  4. Create or assign an appropriate individual compute instance instead of sharing the existing user's session

Explanation

Correct Answer

D:

An AML compute instance interactive development surface is assigned to one user. Check that both users work on separate authorized instances without sharing privileged notebooks.

Incorrect Answers

A:

The compute instance virtual machine size cannot be changed after instance creation; an AML compute instance interactive development surface is assigned to one user.

B:

Restricted compute access and network boundary policy determine remote connectivity options.

C:

Compute instances incur running compute charges unless stopped or scheduled.

Question 37

A data scientist can read an ML workspace but lacks permission to submit training jobs. A production rollout depends on a reproducible result, not just a successful settings update. The next run should prove that the scientist submits a test job without gaining subscription owner privileges. Which next step fixes the reported cause?

  1. Assign an Azure RBAC role allowing ML job creation at the correct workspace scope
  2. Replace the broad assignment with a suitable scoped AML read or monitoring role
  3. Grant the consumer identity the relevant permission on the registry resource as well as required workspace access
  4. Grant its managed identity the relevant Blob data-plane role at the storage scope

Explanation

Correct Answer

A:

Workspace Reader does not grant every job-management action. A successful run shows that the scientist submits a test job without gaining subscription owner privileges.

Incorrect Answers

B:

Least privilege limits access to the operations required at the intended resource scope.

C:

Registry asset authorization and workspace authorization are separate scopes.

D:

That choice is useful when an automated job obtains an Azure resource-management token but receives a 403 on Azure Blob data; workspace Reader does not grant every job-management action.

Question 38

An Azure ML batch job can read workspace metadata but its compute managed identity receives Key Vault HTTP 403 on `get secret` for the approved vault. Network access passes, and the vault uses the Azure RBAC permission model. What is the least-privilege authorization fix?

  1. Assign Storage Blob Data Reader on the vault resource group.
  2. Grant Key Vault Contributor to the deployment author, without changing the job principal.
  3. Assign Key Vault Secrets User to the compute managed identity at the approved vault scope and retest `get secret`.
  4. Rename the Azure ML workspace and rebuild the compute cluster.

Explanation

Correct Answer

C:

For a vault using Azure RBAC, Key Vault Secrets User permits retrieval of secret contents. Granting that role to the caller recorded in the 403 addresses the denied data-plane operation without broad vault administration.

Incorrect Answers

A:

Blob data permissions authorize Azure Storage object reads; they do not grant the Key Vault secret value operation named in the 403 event.

B:

Key Vault Contributor manages vault control-plane resources but does not read secret values. The denied caller is the compute managed identity rather than the author.

D:

Neither a workspace rename nor a new compute resource grants the current principal the missing Key Vault secrets data action.

Question 39

A cross-team engineer needs only permission to monitor model jobs but was assigned subscription Owner. The proposed fix needs to work without changing unrelated approved resources. Before rollout, the engineer must confirm that the engineer sees job results but cannot grant Azure roles. What is the best next action for this Azure AI workload?

  1. Assign an Azure RBAC role allowing ML job creation at the correct workspace scope
  2. Use an approved workload identity federation flow and narrowly scoped Azure role assignment
  3. Grant the consumer identity the relevant permission on the registry resource as well as required workspace access
  4. Replace the broad assignment with a suitable scoped AML read or monitoring role

Explanation

Correct Answer

D:

Least privilege limits access to the operations required at the intended resource scope. The observable result is that the engineer sees job results but cannot grant Azure roles.

Incorrect Answers

A:

Workspace Reader does not grant every job-management action. Least privilege limits access to the operations required at the intended resource scope.

B:

Federated identity replaces stored deployment secrets with controlled short-lived token exchange.

C:

Registry asset authorization and workspace authorization are separate scopes. Least privilege limits access to the operations required at the intended resource scope.

Question 40

A service deployed on AML compute authenticates to Azure resources using a hard-coded user password. The next test must distinguish a real fix from a configuration change that leaves the symptom intact. After remediation, the operator will verify that the service retrieves its secret without storing a username or password. Which next step fixes the reported cause?

  1. Enable an approved managed identity and grant only the required resource-specific permissions
  2. Grant the consumer identity the relevant permission on the registry resource as well as required workspace access
  3. Grant its managed identity the relevant Blob data-plane role at the storage scope
  4. Use a purpose-built least-privilege role with the exact deployment actions required

Explanation

Correct Answer

A:

Managed identities let Azure workloads obtain tokens without distributing permanent user credentials. On retest, the service retrieves its secret without storing a username or password.

Incorrect Answers

B:

Registry asset authorization and workspace authorization are separate scopes; managed identities let Azure workloads obtain tokens without distributing permanent user credentials.

C:

Management-plane role assignments do not necessarily authorize data-plane object reads. But managed identities let Azure workloads obtain tokens without distributing permanent user credentials.

D:

Broad Contributor privileges include unrelated resource modifications that exceed model deployment needs.

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