AWS Machine Learning with Amazon SageMaker AI

Module 1 – Introduction to Machine Learning

• Understand Artificial Intelligence and Machine Learning
• Difference between AI, Machine Learning, and Deep Learning
• Understand models, training, and predictions
• Understand features and labels
• Supervised and unsupervised learning
• Classification, regression, and clustering
• Examples of Machine Learning use cases in organizations

Module 2 – Machine Learning on AWS

• Overview of the AWS Machine Learning ecosystem
• Introduction to Amazon SageMaker AI
• Role of Amazon S3 in a Machine Learning project
• Introduction to IAM roles and permissions
• Introduction to Amazon CloudWatch for monitoring
• Understand the architecture of a Machine Learning solution on AWS

Module 3 – Preparing the AWS Environment

• Access the AWS Management Console
• Select an AWS Region
• Create and configure an Amazon S3 bucket
• Upload a dataset to Amazon S3
• Access Amazon SageMaker AI
• Configure the required permissions
• Explore the SageMaker working environment

Module 4 – Understanding and Preparing Data

• Identify the data used for model training
• Understand rows, columns, and variables
• Identify numerical and categorical variables
• Identify the target variable
• Handle missing values
• Identify and remove duplicates
• Understand outliers
• Select relevant features

Module 5 – Transforming Data for Machine Learning

• Clean a dataset
• Transform data types
• Encode categorical variables
• Normalize data
• Introduction to Feature Engineering
• Create a dataset ready for model training
• Discover data preparation tools in SageMaker

Module 6 – Preparing Training and Test Data

• Understand the purpose of training data
• Understand the purpose of validation data
• Understand the purpose of test data
• Split a dataset into multiple subsets
• Understand model generalization

Module 7 – Training a Model with Amazon SageMaker AI

• Understand how Machine Learning algorithms work
• Select an algorithm appropriate for the problem
• Understand SageMaker Training Jobs
• Configure a training job in SageMaker
• Understand hyperparameters
• Select computing resources
• Launch a training job
• Monitor the training process
• Retrieve the trained model

Module 8 – Evaluating Model Performance

• Understand why models must be evaluated
• Measure Accuracy
• Understand Precision
• Understand Recall
• Discover the F1 Score
• Understand the confusion matrix
• Compare predictions with expected results
• Interpret model performance

Module 9 – Understanding Underfitting and Overfitting

• Understand Underfitting
• Understand Overfitting
• Identify a model that does not generalize well
• Understand how data affects model performance
• Understand the impact of hyperparameters
• Improve a model progressively

Module 10 – Optimizing a Machine Learning Model

• Compare multiple models
• Adjust hyperparameters
• Improve the features used by the model
• Introduction to Hyperparameter Tuning
• Retrain a model
• Compare performance results
• Select the best-performing model

Module 11 – Deploying a Model with Amazon SageMaker AI

• Understand the concept of inference
• Create a deployable model
• Understand SageMaker endpoints
• Configure an endpoint
• Deploy a model
• Send new data to the model
• Retrieve predictions
• Discover real-time inference
• Introduction to batch and serverless inference

Module 12 – Integrating a Model into an AWS Application

• Understand how an application consumes a Machine Learning model
• Call a SageMaker endpoint using Python
• Introduction to AWS Lambda
• Introduction to Amazon API Gateway
• Understand an API Gateway, Lambda, and SageMaker architecture
• Test predictions from an application

Module 13 – Introduction to MLOps on AWS

• Understand the Machine Learning model lifecycle
• Version Machine Learning models
• Introduction to SageMaker Model Registry
• Register different versions of a model
• Understand model approval workflows
• Manage model deployment to production
• Understand the model retraining process

Module 14 – Monitoring a Model in Production

• Introduction to Amazon CloudWatch
• Review endpoint metrics
• Review logs
• Understand Data Drift
• Understand Model Drift
• Introduction to SageMaker Model Monitor
• Identify model performance degradation
• Determine when a model should be retrained

Module 15 – Security and Cost Management

• Understand IAM roles used by SageMaker
• Apply the principle of least privilege
• Control access to data stored in Amazon S3
• Understand data protection
• Identify the main SageMaker cost drivers
• Understand Training Job costs
• Understand endpoint costs
• Remove unused AWS resources after labs

Module 16 – Introduction to Generative AI on AWS

• Understand the difference between traditional Machine Learning and generative AI
• Understand Foundation Models
• Understand Large Language Models
• Introduction to Amazon Bedrock
• Understand the concept of prompting
• Introduction to embeddings
• Introduction to Retrieval-Augmented Generation (RAG)
• Identify use cases for SageMaker AI and Amazon Bedrock

Module 17 – Hands-on Workshop: End-to-End Machine Learning Project

• Import a dataset into Amazon S3
• Explore and prepare the data
• Split the data into training and test datasets
• Configure Amazon SageMaker AI
• Train a Machine Learning model
• Evaluate model performance
• Improve the model
• Deploy the model to a SageMaker endpoint
• Send new data to the deployed model
• Generate and interpret predictions
• Review metrics and logs
• Remove AWS resources used during the workshop

Other course(s) in this category

→ AWS Solutions Architect Associate (SAA-C03) Training – Infrastructure, Serverless, and Certification

→ AWS Machine Learning with Amazon SageMaker AI

→ AWS DevOps Training: Automation and Infrastructure as Code (IaC)