AIF-C01 Question 87
Select 3A company wants to build a generative AI application that can produce high-quality text outputs based on user prompts. They plan to use AWS services to reduce infrastructure management overhead and ensure scalability. Which combination of AWS services and technologies would be most suitable for this use case?
- A
Amazon Bedrock
- B
Amazon SageMaker JumpStart
- C
AWS Lambda for hosting the generative model
- D
Amazon EC2 with custom deep learning frameworks
- E
Amazon S3 for storing training data
- F
AWS CodePipeline for deploying the model
Show answer and explanation
Correct answers: A, B, E
Explanation
To build a scalable generative AI application with reduced infrastructure management, Amazon Bedrock and Amazon SageMaker JumpStart are ideal for leveraging foundation models and pre-trained solutions. Additionally, Amazon S3 supports the storage of training data, providing a scalable and cost-effective foundation for the application. These services work together to simplify the development and deployment of generative AI applications while minimizing operational complexity.
- A. Correct.
Amazon Bedrock is a fully managed service that allows developers to build and scale generative AI applications using foundation models without managing infrastructure. It is suitable for this use case.
- B. Correct.
Amazon SageMaker JumpStart provides pre-built and pre-trained models, including those for generative AI, which simplifies model deployment and reduces time to production.
- C. Incorrect.
AWS Lambda is not ideal for hosting complex generative AI models as it is designed for lightweight, event-driven tasks and has execution time limits.
- D. Incorrect.
Amazon EC2 is a viable option for hosting custom deep learning frameworks, but it requires significant infrastructure management and does not align with the company's goal of reducing overhead.
- E. Correct.
Amazon S3 is commonly used for storing training data and is a scalable and cost-effective solution for this purpose.
- F. Incorrect.
AWS CodePipeline is used for CI/CD pipelines and is not directly relevant to hosting or running a generative AI model.