Professional Cloud ArchitectDemo

By Google · Question Mode

QuestionQ306

Designing for security and compliance

You are designing a new application feature for a healthcare provider located in Qatar. The feature must summarize sensitive patient notes submitted by clinicians. A critical requirement is that the contents of these patient notes must never be processed outside Qatar's borders. You want to use a powerful, pretrained generative model for the summarization task while strictly complying with the data-residency constraint. What should you do?

  • A Write application code that sends the patient notes explicitly to the Gemini API endpoint in Qatar for summarization. Protect the API by VPC-Service Controls.
  • B Use Vertex AI Model Garden to select a Gemma model. Deploy this model to a Vertex AI Endpoint within a Google Cloud region located in Qatar.
  • C Use the Cloud Natural Language API to analyze the text and configure it to generate a summary of the patient notes.
  • D Gather a large, anonymized dataset of medical notes. Use Vertex AI Training to train a custom summarization model from scratch, deploying it in a Qatar region.
Explanation

Deploying a Gemma model from Vertex AI Model Garden to a Vertex AI Endpoint in Doha, Qatar (me-central1) keeps online inference within the selected regional deployment while using a pretrained generative model. Doha supports Vertex AI custom-model online inference, and Model Garden supports deploying Gemma models to Vertex AI endpoints. VPC Service Controls helps mitigate data-exfiltration risks but does not itself establish the geographic location in which a Gemini API request is processed.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!