• 5+ years of experience in building machine learning driven systems. • Hands-on experience with data from retail, military and consumer space industries. • Extracting value from Big Data and tera-scale computing in business settings.  Providing explainable and actionable data insights.  Requirements development in collaboration with domain knowledge experts.  Effective, descriptive, and to the point presentation of experimental results. • Predictive modeling, clustering, classification, and optimization analysis. • Decision trees, random forests, stochastic methods, regression analysis, association rule mining, ensemble classifier learning. • Clustering for pattern discovery in large heterogeneous data sets. • Personalization and recommender systems. • Modeling on high dimensional, sparse, noisy, and incomplete data. • Design of A/B tests across different channels: email, web, and mobile. • Rapid prototyping and deployment (Matlab/R/Python, C/C++, C#/Java, SQL/Hive/Hadoop, HTML5/Django/AngularJS). Often, in parallel and/or distributed processing settings.

Data Scientist @ At livingsocial, I work on a variety of different projects: • Daily email personalization: Everyday livingsocial sends millions of personalized emails to its customers; these emails generate significant amount of revenue of the company. I have proposed and designed a recommendation algorithm, which showed a 46% lift in binary purchase over a baseline algorithm. • Online A/B testing: Based on my analysis, I proposed a web site ( and mobile web) improvement, which lead to an A/B test. During the test, the treatment group showed a statistical significant net revenue lift of 17% and brings $280K per month to the company • Proposed and designed several recommendation diversity metrics for email personalization. These metrics helped us to improve overall quality of the recommendations and increase user engagement. Key technical challenges: high dimensional, sparse, and large volume data, incomplete and noisy data, recommendations with implicit user feedback, time sensitive recommendations. From September 2014 to Present (1 year 2 months) Washington D.C. Metro AreaData Scientist @ At Resonate, my goal is to predict consumers' values, beliefs, and attitudes based on their online behavior. These values, beliefs, and attitudes are the main product of the company, and used for precise consumer targeting. As a member of a research team of 4 PhD scientists, I am responsible for the entire KDD process of all ~4K online behavior data models. Using advanced data preparation procedures and stochastic machine learning (ML) methods, I was able to reduce a typical model training time by 2000 times with the same level of precision. Consequently, with such fast training time, we were able to run more experiments, optimize model parameters, and achieve a higher model quality. I also developed an internal Model Management System that keeps track of a model’s parameters, Machine Learning algorithm used to build the model, its validation results, and other useful meta-data. Key technical challenges: high dimensional, sparse, and large volume data, relevant feature selection, incomplete and noisy data, multiple classes with imbalanced class distribution, high precision predictions. From May 2013 to August 2014 (1 year 4 months) Washington D.C. Metro AreaResearch assistant @ I proposed an Inter-Domain Supervision (IDS) clustering framework to discover clusters within diverse data formats, mixed-type attributes and different sources of data. Results in clustering real data sets with mixed numerical, categorical, visual and text attributes showed that the proposed IDS clustering framework gives improved clustering results compared to conventional baseline methods, in addition to ensemble and multiview clustering, over a wide range of parameters. In my dissertation, I presented a real life application of the IDS approach to the cluster-based automated image annotation problem and presented evaluation results on a benchmark data set, consisting of images described with their visual content along with noisy text descriptions, generated by users on the social media sharing website, Flickr. This work was supported by National Science Foundation. Key technical challenges: Big Data clustering, incomplete and noisy data, high dimensionality and large volumes of data, heterogeneous data of different types and sources. From September 2009 to December 2013 (4 years 4 months) Teaching assistant @ Very Large Scale Integration circuits design Lab. Taught students how to design their own microchip. From August 2009 to December 2009 (5 months) Research assistant @ Computational Intelligence Laboratory. Designed and developed a web-based machine learning repository with a built-in recommendation system. From January 2008 to August 2009 (1 year 8 months) Software Developer, Database Architect @ • Designing and developing ERP, BI (Business Intelligence) and BPM (Business Performance Management) solutions. • Database design From June 2007 to January 2008 (8 months) Software Design Analyst @ Designing and developing ERP systems, starting from an interview and ending with a final product. From May 2006 to May 2007 (1 year 1 month)

PhD, Computer Engineering and Computer Science @ University of Louisville From 2009 to 2013 MS, Computer Engineering and Computer Science @ University of Louisville From 2008 to 2009 MS, Physics @ Perm State University From 2002 to 2007 Artur Abdullin is skilled in: Data Mining, Machine Learning, Algorithms, Pattern Recognition, Big Data, R, Computer Science, Statistics, Data Analysis, C++, MapReduce, Artificial Intelligence, Programming, Matlab, Hive, SQL, Python, Text Mining, Databases, LaTeX, Statistical Modeling, Optimization, Software Engineering, Database Design, Analytics, Computer Vision, C, Mathematical Modeling, Numerical Analysis, Hadoop, Operations Research, Physics, Image Processing, Graph Theory, Social Network Analysis, Applied Mathematics, Natural Language..., Signal Processing, Simulations, Software Design, Weka, Fortran, Parallel Computing, HTML 5, Mathematica, Information Retrieval, Java, Neural Networks, Linux