MLA-C01 exam dumps

MLA-C01 practice question 294 of 458

AWS Certified Machine Learning Engineer - Associate. Associate level, Amazon Web Services. Free question with the correct answer and a full explanation.

MLA-C01 Question 294

Single answer

You are building a machine learning model training pipeline on AWS using Amazon SageMaker. Your team expects workloads to vary significantly between small experiments and large-scale training jobs, depending on project needs. You need to choose the most cost-effective resource allocation strategy. Which approach should you take?

  1. A

    Use on-demand instances for all workloads to avoid over-provisioning and ensure flexibility.

  2. B

    Use provisioned resources to handle all workloads, allowing you to reserve capacity upfront and reduce costs.

  3. C

    Use on-demand instances for smaller workloads and Auto Scaling to handle large-scale training jobs.

  4. D

    Use provisioned resources for smaller workloads and on-demand resources for large-scale training jobs.

Show answer and explanation

Correct answer: C

Explanation

The best approach for a machine learning pipeline with variable workloads is to use on-demand instances for smaller, low-resource tasks and Auto Scaling for large-scale training jobs. This strategy balances cost-efficiency by avoiding over-provisioning while ensuring that large workloads have the necessary resources when needed.

  • A. Incorrect.

    Using on-demand instances for all workloads provides flexibility but may lead to higher costs, especially for large-scale jobs, as you are not leveraging scaling or reserved capacity.

  • B. Incorrect.

    Using provisioned resources for all workloads can reduce costs for predictable workloads, but it is not ideal for workloads with unpredictable or varying demand, leading to potential underutilization.

  • C. Correct.

    Using on-demand instances for smaller workloads and Auto Scaling for large-scale workloads ensures cost-efficiency and flexibility. On-demand instances avoid over-commitment for smaller jobs, while Auto Scaling dynamically adjusts resources for larger jobs, optimizing costs.

  • D. Incorrect.

    Using provisioned resources for smaller workloads is inefficient, as smaller workloads do not require reserved capacity. Similarly, using on-demand resources for large-scale jobs can be more expensive than leveraging Auto Scaling.

Timed practice exam

Take a MLA-C01 practice test under exam conditions

65 questions in 130 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam