Clément T

Adikteev is an advertising company, specialized in app re-targeting. I am part of the research data science team, working on various projects, including: - bidding and pricing, using reinforcement learning - user predictive cohorting for apps through ensemble methods - recommender systems for product feed recommendation - text-mining and deep learning models applied to app segmentation - churn prediction using Recurrent Neural Networks and time series analysis

Deep Learning models for Depth estimation and 3D scenes reconstruction.

Used Gaussian Processes to perform ordinal regression based on the work of Chu and Ghahramani. Made GP-based ordinal regression scalable to large datasets using variational Bayes approximation and GPflow/Tensorflow packages.