A company needs to deploy a generative AI model to edge devices that have limited storage and processing power. The model is large and requires significant resources. They want to maintain accuracy while fitting the model on the devices. Which strategy should they use?

Use a larger edge device with more resources

Use a cloud endpoint and stream data to the edge

Deploy the full model to the edge device

Use model quantization and pruning to reduce the model size

What is the main purpose of orchestrating multiple generative AI models in a solution?

To reduce the total number of API calls

To eliminate the need for prompt engineering

To ensure all models use the same parameters

To combine the strengths of different models to achieve better overall results

A team is deploying a generative AI application and wants to understand how users interact with it and where it fails. They need to capture detailed request traces and user feedback for continuous improvement. Which combination of tools should they use?

Azure Log Analytics for all logs and no feedback

Azure Monitor only for infrastructure metrics

Application Insights for tracing and a custom feedback API

Azure Blob Storage for request logs

A company runs an Azure OpenAI workload with predictable peak hours. They want to minimize costs while ensuring performance during peaks. They have implemented autoscaling, but they notice that scaling takes too long during sudden spikes. Which additional strategy should they use?

Use a larger instance type to handle spikes

Pre-warm capacity by setting a minimum instance count that covers typical peak demand

Disable autoscaling and use a fixed capacity

Set a lower scaling threshold to trigger scaling earlier

Microsoft Certified:Azure AI Engineer Associate

Domain 2Objective 3

Optimize and Operationalize a Generative AI Solution AI-102 Practice Questions (Page 1)

Part of the Implement generative AI solutions domain, which accounts for 15-20% of the AI-102 exam. Microsoft does not publish an official question count, but from its 100-minute exam (~40–65 total, ~6–13 in this domain), expect 2–4 from this objective — we provide 47 practice questions to prepare you well beyond it. (estimate)

47questions here

10free pages

9concepts

15-20%of the exam

Questions 1–5

1foundation · easy
Which Azure service is designed to run containerized AI models on edge devices with limited connectivity?
Select an answer first
2expert · hard
A company needs to deploy a generative AI model to edge devices that have limited storage and processing power. The model is large and requires significant resources. They want to maintain accuracy while fitting the model on the devices. Which strategy should they use?
Select an answer first
3foundation · easy
What is the main purpose of orchestrating multiple generative AI models in a solution?
Select an answer first
4application · medium
A team is deploying a generative AI application and wants to understand how users interact with it and where it fails. They need to capture detailed request traces and user feedback for continuous improvement. Which combination of tools should they use?
Select an answer first
5expert · hard
A company runs an Azure OpenAI workload with predictable peak hours. They want to minimize costs while ensuring performance during peaks. They have implemented autoscaling, but they notice that scaling takes too long during sudden spikes. Which additional strategy should they use?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AI-102

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Microsoft. “AI-102” is a trademark of its owner, used for identification only.