Sample Questions from AWS Certified Machine Learning Engineer – Associate (MLS-C01)

Preview a few questions below — answers are revealed when you take the exam.

  1. A team needs to deploy a machine learning model on AWS using Amazon SageMaker. The model requires high computational power and the ability to scale automatically. What approach should they take to ensure efficient resource utilization and cost-effectiveness?

    • Utilize Amazon SageMaker Studio for model development and deploy it on Amazon EC2 instances with auto-scaling enabled.
    • Deploy the model on Amazon SageMaker hosting services with multi-model endpoints to optimize resource usage.
    • Use AWS Lambda for deploying the model to handle sporadic traffic and reduce costs.
    • Implement the model on Amazon EMR for distributed computing and leverage its auto-scaling capabilities.

  2. Consider the situation where a machine learning model trained on AWS needs to be integrated into a mobile application. Which method solves it best for seamless integration and real-time inference?

    • Export the model to TensorFlow Lite format and use AWS Amplify to integrate it into the mobile app.
    • Deploy the model on Amazon SageMaker and use AWS API Gateway to create an endpoint for the mobile app to call.
    • Convert the model to ONNX format and use AWS IoT Greengrass for edge computing on the mobile device.
    • Utilize Amazon Rekognition for pre-built machine learning capabilities and integrate it directly into the mobile app.

  3. Which option best evaluates the performance of a machine learning model deployed on AWS in terms of accuracy, latency, and cost-effectiveness?

    • Conduct A/B testing using Amazon CloudWatch to monitor metrics and compare results between different model versions.
    • Use AWS Step Functions to orchestrate a workflow that periodically retrains the model and evaluates its performance.
    • Implement a canary release strategy using AWS CodeDeploy to gradually roll out the model and monitor its performance.
    • Leverage AWS Comprehend to analyze text data and evaluate the model's performance based on sentiment analysis results.

  4. How should a professional assess the security implications of deploying a machine learning model on AWS, considering data privacy and compliance requirements?

    • Perform a security audit using AWS Config to ensure that the model complies with industry standards and regulations.
    • Implement encryption at rest and in transit using AWS KMS and AWS CloudHSM to protect sensitive data.
    • Utilize AWS Macie to automatically discover, classify, and protect sensitive data within the model's dataset.
    • Conduct a penetration testing exercise using AWS Security Hub to identify vulnerabilities and mitigate security risks.

  5. What strategy should be applied when optimizing the hyperparameters of a machine learning model deployed on AWS to achieve the best performance?

    • Use Amazon SageMaker Automatic Model Tuning to automate the hyperparameter optimization process.
    • Implement a grid search algorithm using AWS Batch to systematically explore different combinations of hyperparameters.
    • Leverage AWS Systems Manager to centrally manage and configure hyperparameters across multiple model instances.
    • Utilize AWS Step Functions to orchestrate a workflow that iteratively trains and evaluates the model with different hyperparameters.