Practical Data Science with Amazon SageMaker Training Course (NH)

Turning data into actionable insights requires more than theory; you need tools that scale. In this one-day, hands-on course, you’ll learn how to build, train, and deploy machine learning models using Amazon SageMaker. You’ll follow a complete end-to-end data science workflow, from data prep and visualization to model evaluation and tuning, all within the SageMaker platform.

Through a customer churn use case, you'll apply real-world techniques like feature engineering, hyperparameter tuning, and autoscaling. Whether you're a developer or data scientist, this course will strengthen your ability to think critically about model performance and production-readiness using SageMaker’s powerful features.

Course Objectives

By the end of the course, you’ll be able to execute a full machine learning pipeline using Amazon SageMaker. You’ll develop practical skills in model training, tuning, and deployment that apply to real-world business problems.

• Prepare datasets for machine learning with SageMaker
• Train, evaluate, and tune ML models using XGBoost
• Perform hyperparameter tuning with SageMaker tools
• Deploy models to production endpoints with autoscaling
• Analyze model outputs and consider the cost of errors

Agenda

Module 1: Introduction to Machine Learning

• Types of machine learning
• ML job roles and pipeline stages

Module 2: Data Preparation and SageMaker Overview

• Training vs. test datasets
• SageMaker console walkthrough
• Launching Jupyter notebooks

Module 3: Problem Formulation and Dataset Prep

• Business challenge: customer churn
• Exploring and preparing the dataset

Module 4: Data Analysis and Visualization

• Visualizing features and target relationships
• Cleaning and transforming data

Module 5: Training and Evaluating the Model

• Using XGBoost in SageMaker
• Setting up estimators and hyperparameters
• Deploying and evaluating the model

Module 6: Automatic Hyperparameter Tuning

• Creating tuning jobs in SageMaker
• Exercises in parameter optimization

Module 7: Deployment and Production Readiness

• Endpoint deployment
• A/B testing and autoscaling
• Monitoring performance

Module 8: Understanding Cost of Errors

• Error types and business implications
• Adjusting classification thresholds

Module 9: Amazon SageMaker Architecture & Features

• SageMaker in a VPC
• Batch transforms, Ground Truth, Neo