A company wants to deploy a custom-trained model to production. They have strict data residency requirements and need to keep the model and data within their own AWS account. They also want to minimize operational overhead. Which deployment method should they choose?

Use Amazon QuickSight to serve predictions

Deploy the model on an EC2 instance and manage the API yourself

Use Amazon Bedrock to host the custom model

Use Amazon SageMaker to deploy the model as a managed endpoint

A startup wants to add a text summarization feature to their application. They have no in-house ML team and need to minimize infrastructure management. They found a pre-trained open-source model that meets their quality requirements, but they do not want to operate servers or handle scaling. Which approach best satisfies their constraints?

Train a custom model from scratch using Amazon SageMaker

Use Amazon QuickSight to generate summaries

Deploy the open-source model on an Amazon EC2 instance and manage the API themselves

Use Amazon Bedrock to access a managed foundation model API

A company is building a machine learning pipeline for a recommendation system. They have a large dataset that is updated daily. They need to retrain the model regularly to keep it accurate, but they also need to minimize the cost of training. Which approach should they take?

Use incremental learning to update the model with new data without full retraining

Use a smaller dataset for training to reduce cost

Retrain the model only when the data changes significantly

Retrain the model from scratch every day using the full dataset

A company has a deployed ML model that is experiencing data drift. The data science team wants to retrain the model, but they are concerned about the cost of retraining and the potential for the new model to perform worse than the current one. Which approach should they take to manage this risk?

Ignore the data drift and keep the current model

Retrain the model and deploy it immediately to production

Retrain the model, evaluate it on a holdout dataset, and use a canary deployment to gradually roll it out

Retrain the model only when the data drift is severe

AWSCertified AI Practitioner

Domain 1Objective 3

Task Statement 1.3: Describe the AI/ML Development Lifecycle. AIF-C01 Practice Questions (Page 5)

Part of the Content Domain 1: Fundamentals of AI and ML domain, which makes up ~20% of our current practice bank. AWS does not publish an official question count, but from its 90-minute exam (~35–60 total, ~7–12 in this domain), expect 2–4 from this objective — we provide 29 practice questions to prepare you well beyond it. (estimate)

29questions here

6free pages

7concepts

Questions 21–25

21application · medium
A company wants to deploy a custom-trained model to production. They have strict data residency requirements and need to keep the model and data within their own AWS account. They also want to minimize operational overhead. Which deployment method should they choose?
Select an answer first
22application · medium
A startup wants to add a text summarization feature to their application. They have no in-house ML team and need to minimize infrastructure management. They found a pre-trained open-source model that meets their quality requirements, but they do not want to operate servers or handle scaling. Which approach best satisfies their constraints?
Select an answer first
23expert · hard
A company is building a machine learning pipeline for a recommendation system. They have a large dataset that is updated daily. They need to retrain the model regularly to keep it accurate, but they also need to minimize the cost of training. Which approach should they take?
Select an answer first
24expert · hard
A data science team is building a model to detect fraudulent transactions. The dataset is highly imbalanced, with only 1% of transactions being fraudulent. The team's business requirement is to catch as many fraudulent transactions as possible, but they also want to avoid overwhelming the fraud investigation team with false positives. Which metric should they optimize to balance these two goals?
Select an answer first
25expert · hard
A company has a deployed ML model that is experiencing data drift. The data science team wants to retrain the model, but they are concerned about the cost of retraining and the potential for the new model to perform worse than the current one. Which approach should they take to manage this risk?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AIF-C01

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by AWS. “AIF-C01” is a trademark of its owner, used for identification only.

Task Statement 1.3: Describe the AI/ML Development Lifecycle. AIF-C01 Practice Questions (Page 5)

Questions 21–25

A company wants to deploy a custom-trained model to production. They have strict data residency requirements and need to keep the model and data within their own AWS account. They also want to minimize operational overhead. Which deployment method should they choose?

A company is building a machine learning pipeline for a recommendation system. They have a large dataset that is updated daily. They need to retrain the model regularly to keep it accurate, but they also need to minimize the cost of training. Which approach should they take?

A company has a deployed ML model that is experiencing data drift. The data science team wants to retrain the model, but they are concerned about the cost of retraining and the potential for the new model to perform worse than the current one. Which approach should they take to manage this risk?

Keep studying — Content Domain 1: Fundamentals of AI and ML