Certified AWS Machine Learning
Amazon Sagemaker
Prepare data for machine learning
Create model in Jupyter notebook
Train your model, hyperparameter optimization
Deploy and manage models on powerful infrastructure
As a machine learning engineer, who is experienced with AWS, I may use a variety of services to prepare, build, train, tune, and deploy machine learning models.
For data preparation, I may use Amazon S3 for storage and Amazon Glue for data cleaning and transformation.
To build models, I use Amazon SageMaker, a fully managed service that provides a Jupyter notebook interface for building models.
For training and tuning, I use Amazon SageMaker’s built-in algorithms, or bring my own algorithms using Amazon Elastic Container Registry (ECR) or Amazon Elastic Container Service (ECS) to run them on SageMaker’s managed infrastructure.
For hyperparameter optimization, I use Amazon SageMaker’s built-in capabilities or use AWS Step Functions and Amazon SageMaker Automatic Model Tuning to automate the process.
Finally, for deployment, I may use Amazon SageMaker’s built-in hosting capabilities to deploy models to a variety of environments, such as on edge devices, on-premises, or on other cloud services. I may also use AWS Lambda and AWS API Gateway to deploy models as serverless functions and expose them as RESTful APIs.