End-to-End Machine Learning and MLOps with Amazon SageMaker
Duration
2 Day
Level
Intermediate Level
Design and Tailor this course
As per your team needs
Overview
This 16-hour comprehensive program provides a deep dive into the machine learning lifecycle and MLOps best practices using Amazon SageMaker. Participants will learn how to transition from manual experimentation to automated, repeatable, and reliable production workflows. The course covers the entire spectrum of ML engineering, from data preparation and model tuning to orchestration and real-time monitoring, ensuring that ML projects are scalable, secure, and governed.
Audience
- ML Engineers: Seeking to automatethe deployment and management of models.
- Data Scientists: Transitioning from local notebooks to scalable AWS infrastructure.
- DevOps Engineers: Responsible for building CI/CD pipelines for machine learning (MLOps).
- Cloud Architects: Designing secure and governed environments for enterprise AI.
- Software Developers: Integrating ML capabilities into cloud-native applications.
Prerequisites
- Technical Core: Basic understanding of data science or software development.
- Programming: Familiarity with Python programming.
- AWS Knowledge: AWS Cloud Practitioner-level knowledge is a plus.
- ML Basics: Awareness of machine learning concepts (recommended).
Curriculum
Module 1: Introduction to Machine Learning
-
Benefits of ML and various architectural approaches
-
Framing business problems and defining prediction quality
-
Processes, roles, and responsibilities within an ML project team
Module 2: Preparing a Dataset
-
Data analysis and preparation techniques at scale
-
Tooling overview: Amazon SageMaker Studio and Notebooks
-
Hands-On Lab: Data Preparation with SageMaker Data Wrangler
Module 3: Training a Model
-
Algorithm selection and managed training jobs in SageMaker
-
Hands-On Lab: Training a Model with Amazon SageMaker
-
Feature Highlight: Leveraging Amazon CodeWhisperer in Studio Notebooks
Module 4: Evaluating and Tuning a Model
-
Model evaluation metrics and performance analysis
-
Hands-On Lab: Model Tuning and Hyperparameter Optimization (HPO)
Module 5: Deploying a Model
-
Deployment strategies: Real-time vs. Batch
-
Hands-On Lab: Deploying to a Real-Time Endpoint and Generating Predictions
Module 6: Operational Challenges and Responsible ML
-
Understanding Responsible ML (Ethics, Bias, and Fairness)
-
Team collaboration models and automation requirements for MLOps
-
Monitoring and updating models: Testing and deployment patterns
Module 7: Introduction to MLOps
-
The MLOps Maturity Model: From manual processes to full automation
-
People, technology, security, and governance pillars
Module 8: Initial MLOps – Experimentation Environments
-
Bringing MLOps to the research phase
-
Hands-On Lab: Provisioning SageMaker Studio via AWS Service Catalog
-
Demonstration: Lifecycle Configurations for Studio environments
Module 9: Repeatable MLOps – Repositories
-
Managing data for MLOps and version control of ML models
-
Integrating code repositories (Git) into the ML workflow
Module 10: Repeatable MLOps – Orchestration
-
Building ML pipelines for end-to-end orchestration
-
Demonstration: Using SageMaker Pipelines and Projects
-
Hands-On Lab: Automating workflows with AWS Step Functions
Duration
2 Day
Level
Intermediate Level
Design and Tailor this course
As per your team needs
