I apply state-of-the-art machine learning methods to data sets of any kind (images, language, text, users, markets, products etc.).
In order to get the most out the given data this is done in a comparative and complementary way across techniques to benefit from their specific strengths. Once hidden patterns and correlations in the data are learned, predictions are made and tested by measuring their accuracy. Visualization of data, intermediate steps and final results plays an important role, not only for the validation of the algorithms correctness but also for providing a user-friendly and intuitively understandable presentation of the data, the problem and the results.
Machine learning methods include deep learning, classification, prediction, regression, clustering, collaborative filtering, decision trees and some more.
Before applying machine learning to a data set it should be extensively explored, cleaned and preprocessed (data mining). I provide this as a separate service.