Skip to main content

🔥 Data on EKS Is Evolving: Introducing AI on EKS

note

This is a major announcement about the future of this repository and upcoming changes to the structure of our blueprints.

📣 Why the Change?​

The Data on EKS project has grown significantly over time, supporting both Data Analytics and AI/ML communities with production-ready blueprints, patterns, and best practices on Amazon EKS.

As the community has grown and the need for more focused content has become clear, we are splitting this repository into two distinct and specialized projects:

  • 📊 data-on-eks – This repo (you are here!) will continue to provide Data Analytics blueprints and patterns, including Amazon EMR on EKS, Apache Spark, Flink, Kafka, Trino, and other related frameworks.
  • 🤖 ai-on-eks – A new, dedicated repository that will serve AI/ML workloads, including Terraform modules, training pipelines, inference serving, LLM deployment patterns, and performance benchmarks for GPUs.

🗓️ Timeline​

  • Soft Launch of ai-on-eks: KubeCon EU (London), First Week of April 2025
  • Full Migration & Public Launch: By the end of April 2025

We will provide a detailed announcement and URLs during and after the soft launch to help contributors and users easily transition.

🔎 What You Need to Know​

Here’s what this change means for contributors and users:

  1. 🚫 AI/ML Pull Requests (PRs) are currently paused in this repository. Please hold off on submitting new AI-related blueprints or modules until the ai-on-eks repo is publicly available.
  2. ✅ Existing AI blueprints will remain temporarily available in this data-on-eks repository. Once the new ai-on-eks repo is fully launched, all new content and updates will be made there instead.
  3. 📦 Once the migration is complete, updates to existing AI/ML blueprints will no longer be made in the data-on-eks repo. However, we will retain these blueprints here temporarily until all related blogs and documentation have been updated to reference the new ai-on-eks repository.
  4. 🔗 Existing AWS blogs, workshops, and documentation that currently point to the data-on-eks repository will remain unchanged to avoid breaking links. We will raise separate issues to update these references to the new ai-on-eks repository over time, depending on resource availability.
  5. 💬 Going forward, all new issues related to AI on EKS, whether for new content or updates to existing blueprints should be opened in the ai-on-eks repository. We will not address AI on EKS related issues in data-on-eks unless the specific blueprint or content exists only in this repo and hasn’t yet been migrated.

✅ What Stays in This Repo?​

This repository (data-on-eks) will continue to focus on:

  • 📊 Data Analytics Patterns: Spark, Trino, Flink, and other analytics frameworks on EKS

  • 🔁 Streaming Platforms: Kafka, Apache Pulsar, and other event-driven systems

  • 🧩 Job Schedulers: Kubernetes-native job orchestration using tools like Apache Airflow and YuniKorn

  • 🗃️ Distributed Databases: Scalable data stores such as Apache Pinot, ClickHouse, and PostgreSQL on EKS

  • 📦 Terraform Modules: Infrastructure as code templates related to data workloads

  • 🌐 Website & Documentation Content: Including this Docusaurus site

  • 🛠️ Best Practices, Benchmarks & Resources: All content specific to data infrastructure and operations

🚀 What’s Moving to the New ai-on-eks Repository?​

The following components are being migrated to the new ai-on-eks repository (coming soon):

  • 🤖 AI on EKS Section: Including model training, inference, and serving patterns currently hosted under /docs/gen-ai

  • 🛠️ AI Terraform Modules: IaC templates used to deploy end-to-end AI workloads

  • 🧠 Gen AI Patterns: Code for Training and Inference patterns

  • 🛠️ AI-Specific Best Practices: Guidance, architecture recommendations, and scaling patterns tailored to AI/ML use cases on EKS

❤️ Thank You​

We deeply appreciate your feedback, support, and contributions to the Data on EKS project. This change helps us build more focused, scalable, and production-grade solutions for both Data and AI/ML workloads.

Stay tuned for more details as we approach KubeCon EU in April. We can’t wait to share what’s next with you!

— The Data & AI on EKS Maintainers