This Master’s Thesis presents the results obtained during the 6-months internship at Amazon Web Services, a world-known company which provides on-demand cloud computing platforms to individuals, companies and governments, on a paid subscription basis. The main focus of this work is to provide a scientific analysis and a description of the implementation realised to develop a Deep Learning based general purpose recommendation system. After having analysed the state of art for traditional recommendation systems, a few Deep Learning architectures are proposed to solve the problem at hand, and the results obtained on several freely available datasets are compared. Particular attention is given to collaborative filtering methods, based on the history of interaction between user and item, and hybrid recommendation system, which promise to solve some of the shortcoming faced with collaborative filtering methods by enriching user-item interaction with other existing information, in the form of either additional features or textual or graphical input.