A research institute needs to run a large language model for experiments. They have a limited budget and need to scale up to handle occasional large batch jobs, but they also handle sensitive research data that must not be exposed to third parties. Which approach best balances these needs?

Cloud-host the model in a private cloud with dedicated hardware.

Locally host the model on a small cluster and accept longer processing times for large jobs.

Cloud-host the model with a strict data processing agreement.

Use a hybrid approach: local hosting for sensitive data and cloud for non-sensitive data.

Which statement accurately describes a privacy trade-off between cloud-hosted and locally hosted model deployment?

Locally hosted models can provide greater data privacy because sensitive data remains within the organization's own infrastructure.

Cloud-hosted models always provide better privacy because cloud providers have stronger security than any on-premises setup.

Cloud-hosted models are inherently non-compliant with data privacy regulations.

Locally hosted models have no privacy risks because the data never leaves the organization.

Which statement accurately describes a scalability trade-off between cloud-hosted and locally hosted model deployment?

Locally hosted models can scale to any size without additional cost.

Locally hosted models can scale more easily because the organization has full control over the hardware.

Cloud-hosted models can typically scale more easily by adding resources on demand, while locally hosted models require purchasing and installing additional hardware.

Cloud-hosted models have no scalability limitations because the provider has unlimited resources.

An autonomous vehicle system uses an AI model for real-time obstacle detection. The vehicle must make split-second decisions even in areas with unreliable internet connectivity. Which hosting option is most appropriate?

Locally host the model on the vehicle's onboard computer.

Cloud-host the model in a nearby region to minimize network latency.

Use a hybrid approach with cloud for fallback when connectivity is available.

Cloud-host the model with a dedicated low-latency connection.

CiscoAI Technical Practitioner

Domain 1Objective 2

1.2 Compare Model Hosting Options (cloud-Hosted vs Locally Hosted) and Their Trade-Offs (cost, Latency, Privacy, Scalability) AI-TECHNICAL-PRACTITIONER Practice Questions (Page 2)

Part of the Generative AI Models domain, which accounts for 20% of the AI-TECHNICAL-PRACTITIONER exam. Cisco does not publish an official question count, but from its 60-minute exam (~25–40 total, ~5–8 in this domain), expect 1–2 from this objective — we provide 10 practice questions to prepare you well beyond it. (estimate)

10questions here

2free pages

5concepts

20%of the exam

Questions 6–10

6application · medium
A research institute needs to run a large language model for experiments. They have a limited budget and need to scale up to handle occasional large batch jobs, but they also handle sensitive research data that must not be exposed to third parties. Which approach best balances these needs?
Select an answer first
7foundation · easy
Which statement accurately describes a privacy trade-off between cloud-hosted and locally hosted model deployment?
Select an answer first
8foundation · easy
Which statement accurately describes a scalability trade-off between cloud-hosted and locally hosted model deployment?
Select an answer first
9application · medium
An autonomous vehicle system uses an AI model for real-time obstacle detection. The vehicle must make split-second decisions even in areas with unreliable internet connectivity. Which hosting option is most appropriate?
Select an answer first
10foundation · easy
An organization deploys a generative AI model on its own GPU servers in a private data center. Which hosting option does this represent?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

No more pagesBack to AI-TECHNICAL-PRACTITIONER

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 Cisco. “AI-TECHNICAL-PRACTITIONER” is a trademark of its owner, used for identification only.