Cubitic is a UK start-up offering machine learning as a service. The company was founded by three veterans of the gaming industry with a background in building distributed data engineering solutions. They have created a platform that offers an on-demand service to run machine-learning algorithms at scale and support behavioral prediction for mobile applications. Apps can stream events to the platform via the Cubitic API which will execute predictive algorithms on the data received in real-time: Moreover, apps can also consume prediction results in order to tailor user-experience based on likelihood of churn, spend or conversion.
The functionality delivered by the service can also be applied to use cases such as failure rate predictions for Internet of Things devices and rejection rate predictions for manufacturing, where as with mobile apps, connected devices can both contribute to and learn from prediction data generated by platform.
The API also supports batch upload and processing of data in order to act as a seed to train the Cubitic algorithms for a customers dataset. Cubitic will also perform custom algorithm development where customers have specific use cases not supported out-of-the-box.
The Cubitic API is REST-based and supports both the real-time and batch features of the service. Cubitic is currently available as a private beta and you can register here.