AWS Certified AI PractitionerFoundation ModelsMedium
An AI solutions architect is designing a system for a large enterprise that requires a foundation model to summarize legal documents. Due to strict data governance policies, the legal documents cannot leave the company's private network, and low latency is critical for real-time lawyer assistance. Which deployment strategy for the foundation model is most appropriate?
- AOn-premises deployment
- BEdge device deployment
- CCloud-based SaaS API
- DHybrid cloud deployment
Show answer & explanationAnswer & explanation
Correct answer: A. On-premises deployment
On-premises deployment is the most suitable strategy when strict data governance requires data to remain within a private network and low latency is critical, as it keeps both the data and the model computation local.
Why the other options are wrong
- B. Edge device deployment is typically for smaller models and limited compute, not large foundation models for complex legal summarization.
- C. Cloud-based SaaS API would violate data governance policies by sending sensitive data off-network.
- D. Hybrid cloud might expose data or introduce latency if the model or data moves to the public cloud.
On-premises FM Deployment
Deploying and running a foundation model within an organization's own private data center or infrastructure.
- Offers maximum control over data security and compliance
- Provides lowest latency for local data processing
- Requires significant internal IT resources and expertise
Memory trick: Deploy FMs: On-Prem, Cloud, Hybrid, or Edge.