Rebuilding a global B2B API to cover more than £500m in risk

in Agile methodologies, Software development, Team as a service

Client

Our client is one of the world’s leading, innovative iGaming insurance companies. Their scalable platform insures risk management for high-prize lotteries. This is a definitive need as lottery operators always risk players winning, no matter the odds that are in place. What our client does is assume that risk on behalf of their clients, in return for a fee on each insured bet.

The platform provides a B2B API for clients to communicate the game information, and all the insured bets, supporting over 50 different lottery formats.

The Challenge

As the top insurance company that allows players to win up to several millions in prizes, our client was in critical need of a robust and highly scalable platform. They required a team they could trust to develop their application’s APIs and to also maintain its high performance.

By implementing our strong Agile project management methodologies, we quickly assembled the perfect team to develop this project. Our client already had a pre-existing MVP of their platform, with minimal functionality, so Hexis’ first hurdle was managing the project hand-over. This challenge was swiftly overcome, despite the limited technical documentation, which led to a significant amount of reverse engineering work.

Within six months, Hexis had delivered beyond the initial scope. Not only had we implemented robust functionalities to enable new customer acquisitions for the client, but we also made several improvements to the application infrastructure in AWS. In addition, we set up a staging environment and greatly increased the power of the production environment.

Technical overview

Our client’s application exposes a B2B API to their clients so that they can communicate insured bets, and a back-office to manage partner and product configuration. We developed these as a set of microservices, through iterative processes, to support all the required features. All services are deployed in an AWS infrastructure, using the following stack: Java 8, Spring Boot (MVC, Security, Data, Actuator), Angular, Hazelcast, AWS, Elasticsearch, NGINX, Tomcat, GitHub, Docker, ansible and Shell script. For management we use the Atlassian stack: Jira and Confluence.

Hexis approach

Hexis was not only responsible for developing the code for our client. We were instrumental in defining their product’s success requirements.

Using Agile methodologies, we accurately validated the needs of this software development project. This included the implementation of new features, refactoring and constant improvements.

As we were reworking an existing MVP, we also employed SCRUM-based methodology. This enabled our team to prioritize any desired new features alongside enhanced legacy code, in a way that supported the growth of robust software.

Working as a remote team introduces several challenges. Our solution is a well-defined development process to ensure quality:

Analysing the required features with the client / product owner before development.
Analysing the system impact of these features and selecting the best implementation approach.
Effectively communicate all development issues and setbacks with the client and implement the best resolution.
Add automatic tests to insure the solution complies with the requirements.
Always perform code review and have the approval of at least another team member.

With these standards we can mitigate risk and have never needed to deploy an application rollback. This process provides our clients with transparency on what is being developed, avoiding any miscommunication.

Summary of benefits

Our client had complete trust and confidence in the Hexis team and our strong ability to build a high-value, complex product. We quickly developed a robust product, that responded to the client’s significant growing business needs. As a result, Hexis was not only involved in the software engineering but continues sprint planning and feedback processes for future developments.