MLA-C01 Question #73: Real Exam Question with Answer & Explanation

Question

A company is running ML models on premises by using custom Python scripts and proprietary datasets. The company is using PyTorch. The model building requires unique domain knowledge. The company needs to move the models to AWS. Which solution will meet these requirements with the LEAST effort?

Options

• AUse SageMaker built-in algorithms to train the proprietary datasets.
• BUse SageMaker script mode and premade images for ML frameworks.
• CBuild a container on AWS that includes custom packages and a choice of ML frameworks.
• DPurchase similar production models through AWS Marketplace.

Explanation

SageMaker Script Mode lets you bring your existing PyTorch training scripts almost as-is, with minimal changes, while AWS manages the pre-built framework container - making it the lowest-effort migration path for teams already using a supported ML framework.

Why the distractors are wrong:

• A - Built-in algorithms require you to reformulate your problem to fit SageMaker's pre-packaged models, which conflicts with the company's custom scripts and unique domain knowledge.
• C - Building a custom container gives maximum flexibility but requires significant Docker expertise and maintenance overhead - the opposite of least effort.
• D - Purchasing marketplace models abandons the company's proprietary datasets and domain-specific logic entirely, which defeats the purpose.

Memory tip: Think of Script Mode as the "lift-and-shift" of ML - you keep your script, your framework (PyTorch, TensorFlow, etc.), and your data logic, while AWS handles the infrastructure plumbing. If the question mentions existing scripts + a named framework + least effort, Script Mode is almost always the answer.

No community discussion yet for this question.