Gamer Arena minimizes DevOps demands, unlocks generative AI potential with AWS

TutoSartup excerpt from this article:
To maintain a rapid pace of innovation across properties and push the boundaries of emerging technologies, Gamer Arena leverages many solutions from Amazon Web Services (AWS)… As a first step, the team experimented with Amazon SageMaker, a fully managed service to build, train, and deploy machine…

Serving more than 550,000 users globally, Gamer Arena offers a unique ecosystem for video game players to compete and earn rewards. Established in 2020, the company aims to introduce gamers to the advantages of blockchain technology in a fun and exciting way. It hosts a mobile application with more than 20 self-published games, as well as a web platform featuring popular games including PUBG, FIFA, and Clash Royale, all of which support Gamer Arena Utility (GAU) cryptocurrency transactions. Most recently, Gamer Arena introduced an artificial intelligence (AI) red panda generator to its digital currency hub gaulaxy labs. To maintain a rapid pace of innovation across properties and push the boundaries of emerging technologies, Gamer Arena leverages many solutions from Amazon Web Services (AWS).

“We are committed to continuous innovation to provide the best possible experience for our users. Our journey with AWS services has been instrumental in enhancing our operations and engagement, particularly through our Web3 products, gaulaxy.com, and our unique AI-driven mascot. By utilizing AWS, we ensure scalability, reliability, and cutting-edge technology integration,” said Gamer Arena CTO Beyazıt Karınca.

Headquartered in Estonia, with operations in Istanbul, Gamer Arena’s offerings are developed by a dedicated team of ten people. Using AWS services, Gamer Arena minimizes DevOps costs and enables its developers to focus on creating new features, like the red panda generator, instead of infrastructure management.

Gamer Arena team.

Experimenting with generative AI

Already embracing emerging technologies, Gamer Arena quickly recognized the transformative potential of generative AI with the rise of public chatbots and began exploring how it could be used to enhance user engagement, streamline operations, and personalize the gaming experience. As a first step, the team experimented with Amazon SageMaker, a fully managed service to build, train, and deploy machine learning (ML) models.

“Amazon SageMaker made it easy for us to learn about AI as we were getting started, and we found it very helpful for training models and examining existing models,” explained Karınca. “By using popular tools available today, we gained insights into how we could benefit from these tools. This also provided us with a roadmap for the workflows we will implement in future projects.”

The team launched an AI-driven mascot, the red panda image generator, as one of its first gamer-facing AI developments in February 2024. Envisioned to encourage deeper engagement, the tool allows users to create personalized avatars and experiment with prompt engineering, fostering a sense of ownership and creativity. Gamer Arena also used the tool to generate red panda advertisement images, further demonstrating the potential of AI-driven features.

Realizing red panda

To run the red panda image generator, the team developed Stable Diffusion models trained offline and deployed on an Amazon Elastic Compute Cloud (Amazon EC2) instance. This integrated with services running in Amazon Elastic Container Service (Amazon ECS), along with payment methods, previous user images, token rights, and the blockchain connection. The team opted to use Amazon EC2 and Amazon ECS simultaneously so that they could leverage the Python-based Stable Diffusion model in a modular manner. Following tutorials, they completed setup and had the implementation live in one day. As Gamer Arena deepens its implementation of generative AI, it’s evaluating how it could create a more optimized solution using Amazon Bedrock.

“As soon as we became aware of Bedrock, we started exploring it, and the Amazon Titan and Anthropic Claude models have caught our interest,” shared Karınca. “By fine-tuning the models on Bedrock, we will be able offer voice assistant and live support services, and we will also include human-like bots that can chat with people in our metaverse project.”

Generative AI considerations  

To protect data privacy, the Gamer Arena team instated robust data protection measures. They also collaborated with AI experts to ensure the technology integrated seamlessly within the existing site. Preparing for the future and ongoing AI implementations, the company trained in-house staff on AI technical complexities, allowing them to overcome hurdles more easily and fully harness AI’s capabilities.

“Implementing generative AI has revolutionized our workflow,” said Karınca. “It has automated aspects of content creation, allowing our team to focus on more strategic tasks. Personalized recommendations have enhanced user experiences, making our platform more engaging and dynamic. We are excited about the endless possibilities that AI brings to our platform.”

AWS in action

While Gamer Arena uses many AWS services, it relies most heavily on Amazon ECS, which publishes Dockerized projects in a platform-independent manner and can scale them with simple configurations. Additionally, the company uses Amazon Elastic Container Registry (Amazon ECR) to store and back up Docker images quickly, and Amazon EC2 to deploy projects with API calls, with Elastic Load Balancing efficiently distributing traffic across virtual machines.

Amazon Relational Database Service (Amazon RDS) centralizes Gamer Arena data, and Amazon Route 53 handles domain and alias redirections. Games are hosted and in-game images are stored on Amazon Simple Storage Service (Amazon S3), which also provides a secure location for project secret keys. Amazon Simple Notification Service (Amazon SNS) and Amazon Simple Email Service (Amazon SES) send player communications, such as verification emails and mobile notifications.

In terms of generative AI development, access to integrated AWS services allows the Gamer Arena team to work more quickly and efficiently. “The multiple deployment options available with AWS, such as SageMaker, ECS, EC2, and Bedrock, makes it easier to plan how to develop the fastest solution,” noted Karınca.

He concluded, “As a technology startup with predominantly young users, people want to see new things early. Thanks to AWS, we can stay current with these developments. Our vision for the future is ambitious and dynamic. We aim to significantly expand our user base, bringing the thrill of Gamer Arena to a broader audience worldwide. By leveraging the latest in AI and Web3 technologies, we plan to push the boundaries of what’s possible in gaming.”

Learn more about building and scaling generative AI applications with AWS.

Gamer Arena minimizes DevOps demands, unlocks generative AI potential with AWS
Author: Emily McKinzie