We recently introduced our Divante Tech Radar. As a growing company and one that is working in eCommerce, loyalty, and innovations, we find ourselves in contact with more and more technologies. Each is right for a different job, and we see how our various teams favor or rely upon very different solutions; however, we also see that certain technologies are evergreen while others fall out of favor, and we notice emerging technologies that are starting to be more widely adopt and become staples. We’re breaking down the Radar into more detail but discussing how each of our teams interact with the technologies they personally recommend—and asking why they are right for business.
Technology Radar: Open Loyalty Team
The technology stack is always set for a specific purpose and context. These use cases are the most interesting part of the Technology Radar we’ve built at Divante. Today I’m interviewing Cezary Olejarczyk from the Open Loyalty Team at Divante about their technology insights.
Compare all technologies in our live Tech Radar: View it here
PK: Hello Cezary! Please introduce yourself and say something about Open Loyalty and your role?
CO: Hello, I’m Cezary, CEO of Open Loyalty. Open Loyalty is a technology for developing dedicated loyalty program applications. It’s the world’s first open-source loyalty system and is prepared for large-scale projects and proven by partners.
PK: Nice! It looks like you have to deal with huge data sets regarding users’ behavior. That’s the reason you marked ELK Stack with Adopt on the Radar? What’s the exact use case for ELK?
CO: Yes, we collect a massive amount of customers’ behavioral data. And Elasticsearch is a good fit for us. First of all, Elasticsearch is a mature technology with built-in scalability and performance. Secondly, it allows for great reporting and analytics with Kibana. This is a huge advantage that our clients appreciate as they can create customizable reports and use predictive analytics.
PK: This leads us to Kubernetes and Amazon EKS. Can you say something more about your use case for the K8 stack? Why did you decide to use an Amazon-based setup?
CO: Open Loyalty was based on Docker containers and images since day one. Kubernetes is an outstanding tool for managing containerized environments for massive applications. Although dealing on your own with k8s cluster configuration is not an easy thing to do, that is where Amazon EKS comes into play. It is a great compromise between flexibility and ease of usage. It allows you to set up a cluster within minutes, deploy the application, and automatically scale it up and down to keep great performance and costs in balance.
PK: One of the really interesting items you put in the ‘Trial’ ring of the Radar is Azure Cognitive and Keras (AI framework). How do you use this technology in relation to loyalty applications?
CO: At Open Loyalty, we strongly believe that AI and Machine learning can disrupt the loyalty industry. That’s why we started with a quick PoC to implement a seamless loyalty program. The idea is to use facial recognition based on Machine Learning to identify a loyalty member at stores and award them with points for shopping. In order to test this idea, we installed a standard IP camera at the front door of the Divante office and connected it with Azure Cognitive to identify who’s entering the office. Once an employee was identified, he was automatically registered in Open Loyalty and awarded points for entering the office on time. At first, we dealt with a lot of false-positives in the facial recognition. To improve overall installation, we used Keras to create and train our own face recognition model and put it in front of the Azure service.
PK: Wow, that sounds like a Science Fiction. Getting back to the ground I see that there is also a cooperation technique called “Behave” on the radar. How you use it and why is it important?
CO: Behave is a tool for behavior-driven development, backed with Python as a code. It was introduced by our QA team. We use behave to test the behavior of Open Loyalty’s API. Behave uses the Gherkin language Given/When/Then which is easy to understand for business and encourages them to work with technical people.
PK: Regarding Quality Assurance, I see some tools like Behat, Newman, and Sonar Cube in the ‘Trial’ ring. Could you tell us more about how you approach quality assurance and automated tests?
CO: There is always room for improvement. Testing new technologies and tools allows us to asses the best fit for us. If there is something that can improve in our QA or automate testing, then it’s worth at least checking out.
PK: Thank you for your time Cezary!
CO: It’s a pleasure.
Check out the Technology Radar and see how the technology we have discussed here fits into the greater scheme of things and how strongly we recommend each. Look out also for our future conversations with other teams. Each build requires a specific set of skills, people, and applications, so we’ll discuss the choices that each team is making along the way.