I. Overview:

Artificial intelligence and machine learning (AI/ML) are becoming mainstream. In this course, students will spend a day in the life of a data scientist so that students can collaborate efficiently with data scientists and build applications that integrate with ML. Students will learn the basic process data scientists use to develop ML solutions on Amazon Web Services (AWS) with Amazon SageMaker. Students will experience the steps to build, train, and deploy an ML model through instructor-led demonstrations and labs.

II. Duration: 08 hours (1 day)

III. Objectives:

• Discuss the benefits of different types of machine learning for solving business problems
• Describe the typical processes, roles, and responsibilities on a team that builds and deploys ML systems
• Explain how data scientists use AWS tools and ML to solve a common business problem
• Summarize the steps a data scientist takes to prepare data
• Summarize the steps a data scientist takes to train ML models
• Summarize the steps a data scientist takes to evaluate and tune ML models
• Summarize the steps to deploy a model to an endpoint and generate predictions
• Describe the challenges for operationalizing ML models
• Match AWS tools with their ML function

IV. Intended Audience:

• Development Operations (DevOps) engineers
• Application developers

V. Prerequisites:

• AWS Technical Essentials; Entry-level knowledge of Python programming; Entry-level knowledge of statistics.

VI. Course outlines:

1. Module 1: Introduction to Machine Learning

• Overview
• Introduction and Benefits of Machine Learning (ML)
• Types of ML Approaches
• Framing the Business Problem
• Prediction Quality
• Processes, Roles, and Responsibilities for ML projects
• Knowledge Check

2. Module 2: Preparing a Dataset

• Overview
• Data Analysis and Preparation
• Data Preparation Tools
• Demo: Review Amazon SageMaker Studio and Notebooks
• Knowledge Check
• Lab 1: Data Preparation with SageMaker Data Wrangler

3. Module 3: Training a Model

• Overview
• Steps to Train a Model
• Choose an Algorithm
• Train the Model in Amazon SageMaker AI
• Knowledge Check
• Lab 2: Training a Model with Amazon SageMaker

4. Module 4: Evaluating and Tuning a Model

• Overview
• Model Evaluation
• Model Tuning and Hyperparameter Optimization
• Knowledge Check
• Lab 3: Model Tuning and HPO with Amazon SageMaker

5. Module 5: Deploying a Model

• Overview
• Model Deployment
• Knowledge Check
• Lab 4: Deploy a Model to a Real-time Endpoint and Generate a Prediction

6. Module 6: Operational Challenges

• Overview
• Responsible ML
• ML team and MLOps
• Automation
• Monitoring
• Updating Models (Model Testing and Deployment)
• Knowledge Check

7. Module 7: Other Model-Building Tools

• Overview
• Different Tools for Different Skills and Business Needs
• No-code ML with SageMaker Canvas
• Amazon SageMaker Studio Lab
• Demo: Overview of Amazon SageMaker Canvas
• Knowledge Check
• Lab 5: Integrating a Web Application with Amazon SageMaker Model Endpoint