Using AWS enables us to focus more on key business outcomes without dealing with infrastructure pressure. We chose AWS because it provides various third-party products and continues to launch machine learning services.
Bon-jun Koo CTO, ST Unitas

ST Unitas is an EdTech company founded in 2010 with a vision of “removing the gap between the world’s rich and poor through education.” ST Unitas generated around 420 billion KRW (360 million USD) in annual sales in just seven years and has approximately 70 leading brands, including Youngdangi (for standardized English tests such as TOEIC and TOEFL) and Gongdangi (for civil service exams). The company also launched an artificial intelligence (AI) education service called STELLA in 2018, which was designed to help each student overcome their weaknesses, as well as a knowledge platform called CONECTS. With these, ST Unitas is expanding its services from education to knowledge.

Taking over the Princeton Review in 2017, one of the top education institutions in the US, ST Unitas is growing into a global EdTech company. To handle the vast amount of video content and education data, the company is using machine learning on AWS. ST Unitas aims to provide services tailored to its customers’ needs through more organized and stable big data management.

Previously, ST Unitas, which provides user-based operation services, predominantly operated its main system in on-premises data centers. However, as the number of users increased, data center costs also continued to rise. On top of that, the operation of its data centers and servers was time consuming. When the number of users spiked and immediate action was required, they were not in a position to immediately respond. As a result, they needed a flexible infrastructure to deal with the sudden increase in traffic and a solution that could reduce costs. Taking over the Princeton Review, ST Unitas required a cloud vendor with a worldwide network that would allow it to expand as a global EdTech company.

ST Unitas tested and examined various cloud providers prior to choosing Amazon Web Services (AWS). AWS offers a number of cost-optimized solutions with pay-as-you-go pricing, has customer case studies across industries, and continues to launch new services. ST Unitas CTO, Bon-jun Koo, explains, “When we began using AWS in 2015, I came to the conclusion that if we had more experience using the cloud, we would be able to focus more on our key businesses without worrying about infrastructural expenses. AWS, which has a variety of third-party products, continues to launch the machine learning services that ST Unitas needs. The intuitive user interface makes things easy for DevOps, so we decided to move to AWS.” For the services that AWS doesn’t provide, ST Unitas can receive help from Megazone, an APN partner.

ST Unitas migrated around 80% of its services from on premises to the AWS Cloud and kept the rest in data centers, maintaining a hybrid infrastructure. Along with Amazon Relational Database Service (Amazon RDS) and AWS CloudFormation, ST Unitas is currently using Amazon Elastic Container Service for Kubernetes (Amazon EKS), which can deploy, manage, and expand containerized applications. Amazon Kinesis Data Firehose and Amazon Redshift are also being used for data streaming. ST Unitas's “Problem Search Engine” is a feature that recognizes images when users take pictures of a problem. Once it recognizes the image, the engine searches for the solution to the problem. This search engine uses Amazon Rekognition along with the company's own image recognition engine.

For the first time in the Korean education industry, ST Unitas signed an ‘Enterprise Support’ contract with AWS and are currently operating an optimized AWS cloud environment with the help of professional engineers who assist them with planning and building solutions. Mr. Koo explains that, “We cut almost 60% of expenses thanks to this AWS Enterprise Support contract as well as our purchasing of reserved instances.” After migrating to AWS, building an environment for the new project, CONECTS, took just 1/10 of the time that would have taken using the existing data centers.

With the AWS environment, deploying source code, storing and managing big data, and planning and building data architecture is 14 times faster than compared to using on-premises data centers. Bon-jun Koo, CTO, explains, “The advantage of AWS is that it specializes in easily building infrastructure. It is a great service in terms of service operation since it is very flexible to use, and it enables us to easily create and test prototypes.” Since AWS has infrastructure around the world, ST Unitas expects it to be helpful in building infrastructure abroad without worrying about latency between regions.

Beginning as a startup, ST Unitas achieved significant growth based on big data and innovative education technology. ST Unitas plans to move beyond simply being an education provider and become a knowledge platform that uses machine learning on AWS to offer personalized services to its customers. According to the CTO, “With AWS, we were able to enhance both service security and stability, and that has marked a significant milestone in providing quality customer service. In the future, we plan to solidify our work with AWS and work toward becoming a global company.”