Amazon SageMaker for Generative AI Applications
1h 22mAdvanced2025-06-23
Authors
Kesha Williams
Software Engineering Manager, Speaker, Tech Blogger
Course details
Are you looking to upskill as a data scientist or machine learning engineer? This course was designed for you. Join instructor Kesha Williams as she outlines the skills you need to know to build, customize, and deploy foundation models using Amazon SageMaker. Learn how to pretrain models from scratch, leverage advanced techniques for model customization, and deploy models efficiently to meet various performance and cost requirements. The course also covers integrating MLOps practices, enabling continuous integration, monitoring, and automation of machine learning workflows. By the end of this course, you’ll be prepared to optimize large language model inference and manage MLOps with ease.
Learning objectives
Pretrain and customize foundation models from scratch using Amazon SageMaker, leveraging the tools and resources provided for optimal performance.
Deploy foundation models efficiently, managing inference requests while ensuring accuracy, low latency, and cost efficiency.
Integrate MLOps practices with Amazon SageMaker, including setting up CI/CD pipelines, monitoring models in production, and automating workflows.
Access, evaluate, and customize pretrained models from various providers to meet specific use case requirements.
Optimize large language model inference using NVIDIA GPUs to achieve faster and more efficient performance.
Learning objectives
Pretrain and customize foundation models from scratch using Amazon SageMaker, leveraging the tools and resources provided for optimal performance.
Deploy foundation models efficiently, managing inference requests while ensuring accuracy, low latency, and cost efficiency.
Integrate MLOps practices with Amazon SageMaker, including setting up CI/CD pipelines, monitoring models in production, and automating workflows.
Access, evaluate, and customize pretrained models from various providers to meet specific use case requirements.
Optimize large language model inference using NVIDIA GPUs to achieve faster and more efficient performance.
Skills covered
Amazon SageMakerCloud DevelopmentProgramming FoundationsGenerative AIArtificial Intelligence FoundationsCloud ServicesArtificial Intelligence (AI)Cloud ComputingSoftware DevelopmentOne-Off
Concepts
0. Introduction
- 01 - Welcome to Amazon SageMaker
1. Getting Started with Amazon SageMaker Studio
- 02 - Understanding Amazon SageMaker Studio
- 03 - Set up your SageMaker Studio environment
2. Pretraining Foundation Models
- 04 - Understand foundation models
- 05 - Explore tools and resources for pretraining
- 06 - Pretrain a foundation model from scratch
- 07 - Challenge - Pretrain a foundation model
- 08 - Solution - Pretrain a foundation model
3. Customizing Foundation Models
- 09 - Access pretrained models
- 10 - Customize models for specific use cases
- 11 - Challenge - Customize a pretrained model
- 12 - Solution - Customize a pretrained model
4. Deploying Foundation Models
- 13 - Explore deployment strategies for foundation models
- 14 - Manage inference requests
- 15 - Ensure accuracy, latency, and cost efficiency
- 16 - Optimize LLM inference using NVIDIA GPUs
- 17 - Challenge - Deploy a foundation model
- 18 - Solution - Deploy a foundation model
5. Using MLOps
- 19 - Understand MLOps
- 20 - Integrate MLOps with SageMaker
- 21 - Monitor models in production
- 22 - Automate workflows
- 23 - Challenge - Implement MLOps in SageMaker
- 24 - Solution - Implement MLOps in SageMaker
Conclusion
- 25 - Your SageMaker journey