Graduate projects

All our graduates complete a vigorous learning process and have the opportunity to present their findings to leading cooperate industries in emerging markets.


Using Github activity as a signal that technical workers are looking for a new job

This project goes to show the enormous potential of Machine learning and data-crunching. An abstraction over a deceivingly simple concept: workers who are looking for a job get active on GitHub. And on Linkedin. And on Hacker News. Can a machine know when you are going to switch jobs? Or when you are thinking about it? That would be the right time to approach someone with a job offer.

AI-powered highlights - Deep-learning the interestingness in Starcraft 2 videos

The videogame industry is bigger than Hollywood, yet watching game replays is tedious since you have to go through lots of boring stuff to get to the interesting bits. This portfolio project makes machines understand and predict peaks of interest of an audience, also called highlights. In this case, Starcraft is the subject because of the thriving video community it holds. 


Epileptic patients never know when their next seizure is going to come. They fear leaving home unaccompanied. They cannot drive. They cannot cook. Would their life quality increase if a machine could warn them when one is about to hit them? This project reveals a model designed to predict epileptic seizures in patients with astonishingly high accuracy.