Siametric Systems | Enduro FIT

The client was already an established entity in the equine fitness industry with a pre-existing application on both iOS and Android.

The primary objectives were focused on getting more investor funding and growing the application. This was done by adding a number of key features that addressed feedback collected from users, including app usability improvements as well as the ability to main more insights from the workout data being collected. These features would help to show potential investors the value the app can add.

Siametric wanted to begin introducing AI features into the app to create meaningful insights from the data collected by the app. They had already been through a technology selection process and wanted to use TensorFlow with the model running on the mobile device.

The existing codebase had been worked on by multiple developers with backgrounds that didn’t primarily involve iOS or Android development, therefore the codebase, while holding the potential for innovation had some maintainability and structural issues which was slowing developer productivity and needed to be addressed before and during the process of adding new functionality and features to the application.

In order to gain familiarity with the codebase Coderus starting working on some of the smaller bug fixes and improvements before tackling larger features. This enabled the development team to understand the structure of the app, the way it used the Firebase APIs and the underlying data model.

To introduce AI to the Enduro application, Siametric had engaged with a machine learning expert to create and train a model capable of detecting the horses gait using sensor data from the mobile device. It would be Coderus’ responsibility to collect data required by the model from the phone’s sensors, pass it to the algorithm and capture the output for presentation to the user.

To support training of the model, the app was updated to allow exporting of ride data for use in training the machine learning model. Similarly once the model was integrated into the app and being run on real world data its raw output was also made exportable allowing its results to be validated.

Unit tests and Test-Driven Development was used for new features to ensure high quality features were delivered. It also helped ensure existing features were not affected by other code changes and allowed the team to find bugs early before builds were released to the client.

During the development process, our team identified that there was no branch protection enabled on the main git branch. By enabling this we removed the the risk of changes being accidentally lost.

The updates and bug fixes carried on his App have allowed Siametric to improve their core features in order to be able to attract further interest in the product.
Siametric can also now collect feedback from users, delete existing workouts, follow colour coded routes, enable audio announcements, and live tracking, all improving the user experience.