Skip to content

ML Model Development Questions

Practice questions for ML Model Development topic in AWS Certified Machine Learning Engineer - Associate. 52 questions covering this domain.

52 questions12 easy27 medium13 hard
Q1
hard

A hyperparameter tuning job is launching many training runs, and the team wants poorly performing runs stopped early to reduce wasted time and cost. W...

Q2
medium

An organization runs the same ML pipeline repeatedly and wants each successful run to register a new version of the resulting model. Which SageMaker f...

Q3
medium

Before a regulated model is reviewed, the team wants a processing job that computes bias metrics and feature attributions and generates explainability...

Q4
medium

A team is creating a hyperparameter tuning job and wants SageMaker to optimize model quality automatically. Which inputs are most important to define ...

Q5
hard

A review board will not approve a model until the team can show why predictions were made and whether bias metrics were evaluated. Which SageMaker ser...

Q6
hard

A company requires every candidate model to remain blocked from production until an authorized reviewer signs off. Which Model Registry attribute shou...

Q7
easy

A team needs a central place to catalog models for production, manage versions, and control approval status before release. Which SageMaker feature sh...

Q8
medium

A company wants to lower the cost of running many hyperparameter tuning jobs but still use SageMaker Automatic Model Tuning. Which additional capabili...

Q9
easy

A compliance reviewer wants to understand why a model made a prediction and also assess model bias. Which SageMaker service should the team use?

Q10
easy

A data scientist wants SageMaker to run many training jobs and automatically identify the best hyperparameter values for a model. Which SageMaker capa...

Q11
medium

Training loss is unstable, and the team wants a SageMaker feature specifically intended to help debug model convergence during training. Which feature...

Q12
hard

A platform team wants a reproducible way to track each pipeline-generated model with version history and associated metadata over time. Which SageMake...

Q13
medium

A machine learning platform team wants to organize all versions of the same model under a single logical collection. Which Model Registry construct sh...

Q14
medium

Which SageMaker feature lets users train and deploy models from a no-code UI for business users with drag-and-drop workflows powered by Autopilot unde...

Q15
easy

Which SageMaker built-in algorithm is well suited for tabular regression and classification on structured data?

Q16
medium

Which Amazon SageMaker built-in algorithm is designed for univariate and multivariate time-series probabilistic forecasting using recurrent neural net...

Q17
hard

An ML engineer wants Bayesian optimization to be the search strategy for a SageMaker hyperparameter tuning job and to maximize area under ROC curve. W...

Q18
hard

A team needs to fine-tune a foundation model on its private data using a managed serverless approach with bring-your-own-data and built-in evaluation....

Q19
medium

An ML team wants to compare metrics, parameters, and artifacts across many SageMaker training trials in a structured way to choose the best model. Whi...

Q20
easy

Which AWS service offers pretrained foundation models from leading providers for generative AI use cases via a single managed API?

Sign in to see all 52 questions

Create a free account to browse all questions — completely free during our launch phase.