On 29-30 September 2015 the training workshop “Deep Architectures for Vision and Natural Language Processing” was organized by ENGINE team and chaired by prof. Markowska-Kaczmar. Over the sixty years of Artificial Intelligence, deep learning is the most exploding area of research. It was a big interest of our event – more than 70 participants was registered. We invited world recognizable researchers as lecturers, from the best universities in deep learning area: Ruslan Salakhutdinov, Department of Computer Science and Statistics, University of Toronto, Canada,. Roland Memisevic, LISA Deep Learning Lab, University of Montreal and Wojciech Zaremba, New York University, Facebook AI Research. All lectures were recorded and the video will be placed on the Platon platform of scientific television and on YoutTube channel.