The project started in August 2014 by Fernando Nogueira and focused on SMOTE implementation. Together with Guillaume Lemaitre, Dayvid Victor, and Christos Aridas, additional under-sampling and over-sampling methods have been implemented as well as major changes in the API to be fully compatible with scikit-learn.
Refers to GitHub contributors page.