They approached Vinta with a difficult task: the site needed improvements on its features and in its load time. The site was slow due to a number of problems on the codebase, identifying and fixing everything was not trivial and required extensive changes to the codebase.
We can quickly understand a code-base and fix existing bugs. Usually we also take a look at the CI/CD process to ensure lasting results.
Evaluating APIs performance, clarity, and abstraction level, we can provide clear maturity improvements.
Frameworks and methods with positive results are incorporated in the client’s process for future use.
Our expertise in code quality, good standards, and refactoring made Vinta ideal for the job. We worked on improving the quality and performance of the codebase resulting in quicker loads and easier maintenance. These changes were done progressively, always in alternation with work on new features. That's one of the reasons they chose Vinta, as they wanted a team that could make the necessary refactoring while making sure customer requirements were still part of the backlog.
The improvements brought many tangible results to the platform: Stacklist loads quickened, it now had less downtime, and any future maintenance will cost way less. Aside from improving code coverage, quality and performance, Vinta also migrated the platform from a structure where one machine was responsible for running all the services to one running on reliable PaaS products that will effortlessly support Stacklist's scaling.
The client was not used to working on an agile environment, which made it challenging for us at the beginning. Given that, we had to adapt. We changed our ways of receiving inputs from them and how things moved from the backlog, making work much smoother after that. These changes enabled productive discussions about features and everything else that needed to be added to the project.
Aside from our usual stack, one of the features required the system was to recommend some tools for registered people based on other users' stack. We used Google's prediction API for recommendations, which allowed us to implement a recommendation engine quickly.