AIF-C01 Question 99
Select 3A retail company is deploying a generative AI solution for personalized customer recommendations. They are considering using Amazon Bedrock to integrate a foundational model (FM). The company wants to minimize costs while ensuring low-latency responses to enhance customer experience. Which of the following strategies would help achieve their goals?
- A
Leverage a pre-trained foundational model provided by Amazon Bedrock instead of training a custom model.
- B
Deploy the solution in multiple AWS Regions to ensure global coverage and redundancy.
- C
Use token-based pricing to optimize costs by controlling the number of tokens processed in each API request.
- D
Provision higher throughput capacity even if it increases costs to ensure low latency for requests.
- E
Implement caching for frequently requested or common responses to reduce the number of API calls.
Show answer and explanation
Correct answers: A, C, E
Explanation
To balance cost minimization and low latency, the company should use a pre-trained foundational model to avoid the high costs of custom models. Additionally, token-based pricing helps optimize expenses by controlling request size. Implementing caching can further reduce the number of API calls, ensuring both cost-efficiency and responsiveness. Deploying in multiple AWS Regions or provisioning higher throughput capacity would increase costs unnecessarily and is not aligned with the company's objectives.
- A. Correct.
Using a pre-trained foundational model avoids the high computational and financial costs associated with training a custom model, making it a cost-effective strategy.
- B. Incorrect.
Deploying the solution in multiple AWS Regions increases availability and redundancy but also significantly increases costs, which contradicts the company's goal of minimizing expenses.
- C. Correct.
Token-based pricing allows the company to control costs by optimizing the size and complexity of requests, aligning with the cost minimization objective.
- D. Incorrect.
Provisioning higher throughput capacity ensures low latency but increases costs, which conflicts with the company's goal of keeping expenses low.
- E. Correct.
Implementing caching reduces the number of API calls, thereby lowering operational costs while maintaining low latency for frequent requests.