Lilly Tech Systems
AWS AI/ML Infrastructure
Build production AI infrastructure on Amazon Web Services. Master EC2 GPU instances for training, S3 data lake architecture, VPC networking for distributed workloads, IAM security policies for ML, and AWS-specific best practices for operating AI at scale.
What You'll Learn
Deep dive into AWS services and configurations for AI workloads.
EC2 for ML
P5, P4d, G5, G6, Inf2, and Trn1 instances: selection, configuration, and optimization.
S3 Data Lake
Building ML data lakes with S3, Glue, and Lake Formation for training data management.
VPC & IAM
Network architecture and security policies tailored for AI/ML workloads on AWS.
Best Practices
AWS Well-Architected ML lens, cost optimization, and operational excellence patterns.
Course Lessons
1. Introduction
Overview of the AWS AI/ML ecosystem, service categories, and infrastructure planning.
2. EC2 for ML
GPU instance families, EFA networking, placement groups, and AMI configuration.
3. S3 Data Lake
S3 bucket architecture, data formats, Glue catalog, and high-performance access patterns.
4. VPC Setup
VPC design for ML workloads, subnets, security groups, VPC endpoints, and EFA.
5. IAM
IAM roles, policies, and permission boundaries for SageMaker, EC2, and S3 ML access.
6. Best Practices
AWS Well-Architected ML practices, cost optimization, monitoring, and operations.
Prerequisites
- AWS account with appropriate permissions
- Experience with core AWS services (EC2, S3, VPC, IAM)
- Basic understanding of ML training and inference
- Familiarity with AWS CLI and CloudFormation/Terraform