Skip to content

Developer Self Service Access to DeepSeek on Amazon EKS

A few weeks back, Tiago Reichert from AWS published a very interesting blog on AWS Community showcasing how you can deploy and use the DeepSeek-R1 LLM on an Amazon EKS Cluster operating in Auto Mode. Detailed step-by-step instructions for this are documented in this Git Repo.

In this blog, we will describe how we took AWS's excellent blog and packaged it to provide a turnkey, 1-click self-service experience for non AWS administrator type users in a typical enterprise. It took one of our solution architects 30 minutes to wrap AWS's example code using Rafay's Environment Manager and PaaS.

Over the last few weeks, we have been asked to demonstrate this every day to several customers and partners. Given the significant interest in DeepSeek and the self service experience, we believe others will benefit from this blog.

Rafay, DeepSeek and EKS

Why Developer Self Service?

The steps documented in the AWS blog may be straightforward for an AWS administrator who may have privileged access to their AWS account. What we hear from our customers is that these steps are not friendly for the typical app developer or enterprise user that is interested in taking DeepSeek for a spin and evaluate it for fit. Here is the the experience they want to provide their app developers etc:

  1. Should not require privileged access to the enterprise's AWS account
  2. Should not be required to learn or use IaC such as Terraform etc
  3. Should not be required to be an expert in AWS
  4. Should not be required to understand how to securely configure and deploy complex infrastructure resources.
  5. Should not expose the DeepSeek endpoints on the Internet or even worse, expose the EKS cluster's API server on the Internet

App Developers should ideally be provided with a menu of standardized options with a 1-click experience to deploy and securely access/use the resources.

What is the End User Experience Like with Rafay?

Step 1

The end user logs into their self service portal (aka Rafay's Developer Hub) and is presented with a menu of standardized options they can select from to use via self service. These options are typically curated by the organization's platform team.

Step 1

Step 2

The user provides a name and clicks deploy. That's it!

Step 2

There is a ton that happens behind the scenes. But, this is abstracted away from the end user. The image below shows a good illustration showcasing the connection between the end user SKU/profile and the backing environment template based on AWS's example code.

Architecture

Info

Click on the link below if you are interested in watching a 2 minute video that showcases the end user experience providing self service creation and use of DeepSeek-R1 deployed on an Amazon EKS Cluster operating in Auto Mode.

About DeepSeek

DeepSeek R1 is an open-weight AI model developed by DeepSeek, designed for advanced reasoning, coding, and language understanding. It is trained on extensive datasets with high-performance computing infrastructure. It is comparable with top-tier models in efficiency and accuracy. Users find DeepSeek disruptive because it challenges established AI players by offering high-performance, open-weight models that rival proprietary alternatives.

  1. Open-Source Alternative – DeepSeek provides open-access models, allowing businesses and developers to fine-tune and deploy AI without vendor lock-in, challenging closed-source giants like OpenAI and Google DeepMind.
  2. Cost-Efficiency & Accessibility – By making powerful AI freely available, DeepSeek lowers the barrier to entry for enterprises, startups, and researchers, disrupting the AI monetization model.
  3. Multilingual Optimization – Unlike many Western-centric models, DeepSeek is optimized for both Chinese and English, making it a strong contender in Asian markets.
  4. Coding & Reasoning Strength – With a focus on programming and advanced reasoning, it directly competes with models like GPT-4 and Claude, appealing to developers and enterprises seeking AI-assisted coding.
  5. Scalable & Efficient AI – Its architecture prioritizes scalability and efficiency, making it suitable for high-performance applications, reducing reliance on expensive, compute-heavy closed AI solutions.

Given these benefits, no wonder there are so many organizations interested in exploring and understanding how DeepSeek can help them.

Conclusion

In this blog, we learnt how easy it is for a platform team to take existing automation, encapsulate it with Rafay Environment Manager and PaaS to deliver a curated and streamlined experience for their end users such as app developers and data scientists.