Customer Stories / Software & Internet

2023

Gemsouls Powers New Virtual Social Experience Globally with Cloud-Native Artificial Intelligence

Gemsouls built a complete data warehouse architecture with Amazon RedShift in two months, and introduced Amazon Rekognition into its social platform for automatic image and video analysis.

1 million

built an architecture for one million users from scratch in a short time

 

2 months

built a complete data warehouse architecture on Amazon RedShift, far more efficient than the six-month self-building plan

95%

unsafe content are automatically flagged

Efficient and cost-effective

accelerated the algorithm R&D process and saved on resources

Overview

Gemsouls is an AI social platform where virtual characters interact with humans using natural languages, and form long-term emotional connections. 

Gemsouls chose Amazon Lightsail VPS instances, together with Elastic Load Balancing and Amazon Route 53, to provide cost-effective back-end support for massively deployed applications. By referring to AWS best practices, and integrating Amazon Simple Storage Service (Amazon S3) and Amazon Identity and Access Management (IAM), Gemsouls has achieved strict internal oversight and external security access control to protect customer data and information. 

Amazon Web Services Case Study: Gemsouls

Opportunity | Delivering A New Social Experience in The Metaverse

Highly anthropomorphic virtual characters and artificial intelligence (AI)-powered virtual social networking services have greatly vitalized the metaverse. Gemsouls, started in 2020, is an AI social platform featuring virtual characters. Using its extensive experience in AI and social networking services, and applying innovative and cutting-edge language models, Gemsouls enables virtual characters to interact with natural languages, to form long-term emotional connections with users.

Gemsouls aims to give everyone an interactive virtual identity in the metaverse, and shape these virtual characters with their own unique habits, personalities, emotions, and even values. It utilizes AI technologies like natural language processing (NLP) and reinforcement learning (RL), and combines user generated content (UGC) and AI generated content (AIGC). Gemsouls also supports the virtual characters that continuously learn and evolve on their own through user interactions.

kr_quotemark

We built a complete data warehouse architecture on Amazon RedShift in two months, instead of a planned self-built one that may take six months. Relying on the analytical capacity of a cloud data warehouse, we can calculate penetration and retention rates of new features faster, which better supports the decisions for product development and iteration, and enhances user satisfaction with our products.”

Gong Qucheng
co-founder & CEO, Gemsouls

Solution | Scaling Infrastructure Architecure in Security and Protection

Gemsouls chose Amazon Lightsail virtual private server (VPS) instances, together with Elastic Load Balancing and Amazon Route 53, to provide cost-effective back-end support for massively deployed applications, and to accommodate the hike in registered and active users. As Amazon Lightsail provides pre-configured, one-click-to-launch resources and a simple integrated load balancing function, Gemsouls launched its business plans quickly without worrying about the high traffic loads. Its technical team can move Amazon Lightsail to Amazon Elastic Compute Cloud (Amazon EC2) with a simple wizard feature to support the growing business, and quickly stand up an architecture for one million users from scratch.

By following AWS best practices, and integrating the security and access control built-in object storage service of Amazon Simple Storage Service (Amazon S3) with granular authority management of Amazon Identity and Access Management (IAM), Gemsouls has achieved strict internal oversight and external security access control, which ensures better protection of customer data and information. This helps it meet the strict requirements of security and privacy protection on its virtual social platform.

Protecting Against Malicious Content

To manage content security and guard against malicious or offensive content, Gemsouls runs Amazon Rekognition on its social platform to automate image and video analysis, letting its technical team screen out inappropriate or offensive contents using fully hosted APIs and customizable review rules. This helps improve compliance and user experience. Amazon Rekognition is able to automatically flag up to 95 percent of unsafe contents, protecting the users and significantly reducing the manual workload.

Adopting Amazon Rekognition has saved Gemsouls significant costs in human resources and time. If it had gone ahead to build its own algorithms, it would have required at least one dedicated engineer who needed to spend two to three months on the project. Using Amazon Rekognition has also removed upfront capex and usage commitments for computer vision technology as Gemsouls needs only pay based on what it uses, further optimizing its business costs.

Gemsouls will deploy the virtual social platform in more countries and regions on AWS global infrastructure. It plans to use Amazon Trainium, AWS second-generation machine learning (ML) chip for deep learning, to develop more smart scenarios and create more diverse and immersive virtual social experience with advanced natural language processing and AI technologies.

Outcome | Optimizing A Data-driven Business For Insights

Gemsouls wants to analyze product operation data, such as user retention and daily active users (DAU), to optimize business decisions and operations. Gemsouls previously analysed its data with MySQL, but it took minutes to run a simple MySQL query from hundreds of millions of data entries. To solve for performance challenges with MySQL, Gemsouls completely migrated the business data to Amazon Redshift cloud data warehouse, and turned to Amazon Managed Workflows for Apache Airflow (MWAA) to read data streams from multiple sources. Now Gemsouls can complete an SQL query in seconds, collect and analyze data efficiently, and drive better business operation and optimization of its products and services.

About Gemsouls

Gemsouls is an artificial intelligence platform powering virtual characters and their connections to the real world, creating a new way for fans to enjoy their favourite characters, and for creators to bring fictional beings to life.

AWS Services Used

Amazon Lightsail

Amazon Lightsail offers easy-to-use virtual private server (VPS) instances, containers, storage, databases, and more at a cost-effective monthly price.

Learn more>>

Amazon Redshift

Amazon Redshift uses SQL to analyze structured and semi-structured data across data warehouses, operational databases, and data lakes, using AWS-designed hardware and machine learning to deliver the best price performance at any scale.

Learn more>>

Amazon MWAA

Amazon Managed Workflows for Apache Airflow (MWAA) orchestrates your workflows using Directed Acyclic Graphs (DAGs) written in Python. 

Learn more>>

Amazon Rekognition

Amazon Rekognition offers pre-trained and customizable computer vision (CV) capabilities to extract information and insights from your images and videos.

Learn more>>

Get Started

Organizations of all sizes across all industries are transforming their businesses and delivering on their missions every day using AWS. Contact our experts and start your own AWS journey today.