Logo Passei Direto
Buscar
Material
páginas com resultados encontrados.
páginas com resultados encontrados.

Prévia do material em texto

Download Valid Microsoft AI-200 PDF Questions with Answers to Study
1 / 6
Exam : AI-200
Title :
https://www.passcert.com/AI-200.html
Developing AI Cloud
Solutions on Azure (beta)
Download Valid Microsoft AI-200 PDF Questions with Answers to Study
2 / 6
1.An AI application needs to access a SQL database. To improve security, you want to eliminate the use
of passwords entirely in the application's connection string.
What should you implement?
A. Shared Access Signatures
B. IP Whitelisting
C. Azure Key Vault Secrets
D. Microsoft Entra ID (Azure AD) Authentication with Managed Identity
Answer: D
Explanation:
Using Managed Identity with Entra ID authentication allows the application to connect to Azure SQL
without any credentials in the code or configuration files.
2.An AI workload on AKS is failing to pull an image from Azure Container Registry (ACR). You verify the
Managed Identity has 'AcrPull' permissions.
What is the most likely networking issue?
A. The ACR has its firewall enabled and doesn't allow the AKS VNet
B. The AKS cluster is in the wrong region
C. The Dockerfile is missing an ENTRYPOINT
D. The pod needs more CPU
Answer: A
Explanation:
If the ACR firewall is enabled, you must either allow the AKS VNet/Subnet or use a Private Endpoint to
allow the cluster to reach the registry securely.
3.An AI application on AKS needs to store a client secret for a 3rd party API. You want to use 'Azure Key
Vault' but you want the secret to appear as a 'File' inside the pod's filesystem.
Which component do you need?
A. Helm Chart
B. Istio Service Mesh
C. Azure Disk Driver
D. Secrets Store CSI Driver
Answer: D
Explanation:
The Secrets Store CSI Driver allows you to mount secrets from Key Vault into a Kubernetes pod as a local
volume (file), providing a secure and familiar developer experience.
4.An AI agent on AKS is running out of memory during peak processing. You want to automatically move
the pod to a node with more RAM.
Which Kubernetes feature handles this?
A. ReplicaSet
B. Service Account
C. Horizontal Pod Autoscaler (HPA)
D. Vertical Pod Autoscaler (VPA)
Answer: D
Download Valid Microsoft AI-200 PDF Questions with Answers to Study
3 / 6
Explanation:
Vertical Pod Autoscaler (VPA) automatically adjusts the CPU and memory 'requests' for your pods,
allowing them to scale up in size (vertically) on larger nodes as needed.
5.An AI agent needs to store the 'Context' of a conversation (the last 10 messages). You want to use a
NoSQL database that offers sub-10ms latency for global users.
Which service should you choose?
A. Azure Cosmos DB
B. Azure SQL
C. Azure Data Lake
D. Azure Archive Storage
Answer: A
Explanation:
Azure Cosmos DB is designed for high-performance, globally distributed applications, providing the
low-latency document storage required for maintaining AI agent state.
6.An AI application uses Azure OpenAI. You need to ensure that the development team can only deploy
'GPT-4o' models and are blocked from deploying older models to save costs.
Which Azure service allows you to enforce this across the subscription?
A. Azure Policy
B. Azure Blueprints
C. Azure Advisor
D. Microsoft Defender for Cloud
Answer: A
Explanation:
Azure Policy can be configured with a 'Deny' effect to prevent the creation of specific resource types or
SKUs (like older model versions) that don't meet organizational standards.
7.An AI developer needs to create a deployment of 'GPT-4o' that has a guaranteed throughput (Tokens
Per Minute) and doesn't fluctuate based on regional demand.
Which deployment type should they choose?
A. Serverless
B. Standard (Pay-as-you-go)
C. Provisioned Throughput (PTU)
D. Global Standard
Answer: C
Explanation:
Provisioned Throughput (PTU) offers reserved processing capacity with consistent latency and
throughput, which is essential for enterprise-grade production AI apps.
8.An AI agent needs to access an Azure SQL Database. To comply with the 'Zero Trust' security model,
you want to use an identity that is automatically managed by Azure and doesn't require a password in the
connection string.
What should you use?
Download Valid Microsoft AI-200 PDF Questions with Answers to Study
4 / 6
A. Shared Access Signature
B. Root Credentials
C. Managed Identity
D. Service Principal
Answer: C
Explanation:
Managed Identities allow the AI agent to authenticate to Azure SQL using Microsoft Entra ID (formerly
Azure AD), removing the need for developers to manage or rotate database passwords.
9.An AI application requires a high-performance vector database. You decide to use the pgvector
extension on Azure Database for PostgreSQL. You want to improve search speed for a large dataset
using an index that builds a graph of neighboring vectors.
Which index type should you use?
A. GIN (Generalized Inverted Index)
B. IVF (Inverted File Index)
C. HNSW (Hierarchical Navigable Small World)
D. B-Tree
Answer: C
Explanation:
HNSW is a graph-based indexing algorithm that provides high-speed, high-recall approximate nearest
neighbor search, ideal for large-scale AI vector data.
10.An AI application on AKS needs a persistent volume that can be mounted by multiple pods
simultaneously for reading and writing data (RWX).
Which storage service is appropriate?
A. Azure Files
B. Azure Disk
C. Local SSD
D. Azure Blob Storage (Archive)
Answer: A
Explanation:
Azure Files (SMB/NFS) is the only managed storage service in Azure that supports the ReadWriteMany
(RWX) access mode required for multiple pods to share a volume.
11.An AI application on AKS needs to access an Azure SQL Database. You have enabled 'Workload
Identity' on the cluster.
Which Kubernetes object represents the identity the pod will use to authenticate?
A. ServiceAccount
B. Deployment
C. ConfigMap
D. Secret
Answer: A
Explanation:
In a Workload Identity setup, the Kubernetes ServiceAccount is federated with an Azure Managed Identity,
Download Valid Microsoft AI-200 PDF Questions with Answers to Study
5 / 6
allowing the pod to authenticate to Azure resources.
12.An AI application requires a globally distributed database with single-digit millisecond latency for
storing user session vectors. You choose Azure Cosmos DB for NoSQL.
Which feature should you enable to ensure the application automatically reads from the nearest region?
A. Service-managed failover
B. Multi-region writes
C. Availability Zones
D. Integrated Cache
Answer: B
Explanation:
Multi-region writes and global distribution in Cosmos DB allow applications to interact with the nearest
local replica for low-latency reads and writes.
13.An AI workload on AKS requires a large amount of temporary disk space (500GB) for processing video
files. You want the fastest possible disk I/O and do not need the data to persist if the node is deleted.
Which disk type should you use?
A. Azure Managed Disks (Standard HDD)
B. Ephemeral OS disk or Local SSD
C. Azure Archive Storage
D. Azure Managed Disks (Premium SSD)
Answer: B
Explanation:
Ephemeral disks and local SSDs provide the lowest latency and highest IOPS because the data is stored
locally on the VM host, which is ideal for temporary AI processing.
14.An AI workload running on AKS requires GPU acceleration for specialized inference tasks. You need to
ensure the pods are scheduled only on nodes with NVIDIA GPUs.
What should you use in the pod manifest?
A. Resource Limits
B. Node Selectors
C. Readiness Probes
D. Taints and Tolerations
Answer: B
Explanation:
Node Selectors (or Node Affinity) allow you to constrain a pod to only run on nodes with specific labels,
such as those indicating GPU availability.
15.An AI application uses Azure Functions to process high volumes of telemetry data. You notice the
function is being throttled. You want the function to scale based on the number of messages waiting in an
Azure Queue.
Which scaling component is used by the platform?
A. The Load Balancer
B. The Traffic Manager
Download Valid Microsoft AI-200 PDF Questions with Answers to Study
6 / 6
C. The Scale Controller
D. The App Service Plan
Answer:C
Explanation:
In the Consumption and Premium plans, the Scale Controller monitors the rate of events (like queue
depth) and determines whether to add or remove function instances.
16.An AI agent needs to access sensitive tenant data in a shared SQL database. You want to ensure that
each tenant can only see their own rows of data, even if the AI agent is compromised.
Which database feature provides this row-level isolation?
A. Transparent Data Encryption (TDE)
B. Dynamic Data Masking
C. Always Encrypted
D. Row-Level Security (RLS)
Answer: D
Explanation:
Row-Level Security (RLS) allows the database engine to automatically filter results based on the identity
of the user or agent, providing strong multi-tenant data isolation.
17.An AI workload on AKS is reporting high memory usage. You want to see which specific pod is
consuming the most resources right now.
Which command provides this real-time resource usage?
A. kubectl top pods
B. kubectl logs
C. kubectl describe pods
D. kubectl get pods
Answer: A
Explanation:
The 'kubectl top pods' command displays current CPU and memory usage for all pods in a namespace,
helping identify resource-intensive AI models.
18.An AI application on AKS needs to access secrets from Key Vault. You want to mount these secrets as
a 'Volume' so the app can read them as files.
Which component is required?
A. Secrets Store CSI Driver
B. Nginx Ingress
C. Helm
D. Azure Disk Driver
Answer: A
Explanation:
The Secrets Store CSI Driver for Azure Key Vault allows you to mount secrets, keys, and certificates
directly into Kubernetes pods as volumes.

Mais conteúdos dessa disciplina