Improving Content Recommendation with Deep Neural Networks

Project Description

Background: 
Taboola is widely recognized as the world’s leading content discovery platform, reaching 1B unique visitors and serving over 360 billion recommendations every month. Recent ComScore data shows that Taboola is second only to Facebook in terms of reach (https://www.taboola.com/press-release/taboola-crosses-one-billion-user-mark-second-only-facebook-world%E2%80%99s-largest-discovery). 
Publishers, marketers, and agencies leverage Taboola to retain users on their site, monetize their traffic and distribute their content to drive high quality audiences. Publishers using Taboola include USA Today, NYTimes, TMZ, Politico.com, BusinessInsider, CafeMom, Billboard.com, Fox Television, Weather.com, Examiner, and many more. 
Taboola's operation is vast with ~2,000 servers in 6 data centers processing big data about users and user behavior, content, pages etc..

What are we looking for?
Taboola is interested in utilizing deep neural networks to improve its predictive capabilities in a number of applications. We are looking for researchers with knowledge and expertise relevant to the use of deep learning for recommendation, Natural Language Processing (inc. speech), and computer vision. More so than looking for experts in certain domains, we are looking for people who have experience in utilizing DL techniques to solve new real world problems, especially where no truth sets exists. These people need to have experience (or at least knowledge) in the various types of network architectures and in particular, use of embedding techniques (like word2vec) for various types of entities (we do that for users, for instance), RNN (LSTM, in particular) and CNN (which are particularly useful for working with images). 
The following papers are very relevant for our inquiry: 
http://arxiv.org/abs/1606.07792
http://arxiv.org/pdf/1607.07326.pdf
http://arxiv.org/pdf/1301.3781.pdf