CLF-C02 ยท AI, machine learning and analytics
SageMaker AI and Amazon Q
Amazon SageMaker AI for building, training and deploying models, and Amazon Q as a generative AI assistant for business and developers.
Key points
- Amazon SageMaker AI (before December 2024: Amazon SageMaker): fully managed service to build, train and deploy your own machine learning models.
- Canvas: no-code models. Ground Truth: data labeling.
- JumpStart: pretrained models. Model Monitor: drift in production.
- Amazon Q Business: generative AI answers from company data.
- Amazon Q Developer: generative AI help with code and AWS operations.
Source: https://docs.aws.amazon.com/sagemaker/latest/dg/whatis.html
Practice questions
Which AWS service is a fully managed platform for building, training and deploying custom machine learning models?
- Amazon SageMaker AI
- Amazon Lex
- Amazon Athena
- AWS Glue
Show answer
Answer: A. Amazon SageMaker AI
Amazon SageMaker AI covers the whole custom-model lifecycle on managed infrastructure. Amazon Lex builds chatbots, Athena runs SQL on S3 and Glue does ETL; none of them trains your own models.
Which tasks can a team carry out with Amazon SageMaker AI?
- Train a model on managed compute without running servers
- Host a trained model on an endpoint for predictions
- Run ad hoc SQL queries directly against files in S3
- Create labeled training datasets with human workers
- Translate product manuals into several languages
Show answer
Answer: A. Train a model on managed compute without running servers, B. Host a trained model on an endpoint for predictions, D. Create labeled training datasets with human workers
SageMaker AI trains models on managed compute, hosts them on endpoints for inference and, through Ground Truth, builds labeled training datasets. Running SQL on S3 is Athena's job and translating documents belongs to Amazon Translate.
To train a model in Amazon SageMaker AI, you must first launch, patch and manage your own fleet of EC2 instances.
Show answer
Answer: False
False. SageMaker AI is fully managed: you pick the training configuration and it provisions and releases the infrastructure for you. Managing your own instances is exactly the heavy lifting it removes.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
