Fullstack MLOps AWS

Purpose of this project is to provide a template to setting up an MLOps stack on AWS EKS using mostly open source available tools.

This repo focuses purely on setting up the infrastructure backbone for AI/ML/LLM applications, which you can then tailor to your own business logic and use case.

Tools

• MLflow - Model Registry
• MLflow - Experiment Tracker
• LakeFS - Data Versioning
• ZenML - Pipeline Orchestrator
• MySQL - RDS backend for MLFlow and ZenML
• S3 - Artifact Storage for ZenML and MLflow
• BentoML/KServe (Still deciding) - Model Deployment
• Terraform - Infrastructure as Code (IaC)
• Terragrunt - Terraform Configuration Management

📚 Documentation Index

The Fullstack MLOps AWS Setup!

Please follow the guides in this order:

1. 0 - Prerequisites
2. 1 - Infrastructure Setup