Worked with the R&D team towards implementing Multi-Arm Bandit based models for solving the problems of bid scaling and impression recommendation in the digital advertising domain.
Researched on the problem of feature selection on a data-set of Sparse Vectors. Developed and implemented feature selection techniques for customer conversion data in digital advertising domain with an aim to identify a certain specific set of features for various campaigns that can be leveraged for better consumer targeting and also save a lot of resources on data-storage while make the prediction platforms run faster. Learned and implemented various algorithms based on Filter and Wrapper Methods like Anova, Fisher Score, Mutual Information, Bayes Error, Kullback–Leibler divergence, Chi-square, etc. Deployed the systems for test on Tera-bytes of data and optimized performance.