Data Scientist

Sanjar Ahmadov

**** ********** ** ****** *****, MI 48326 • 646 – 361 – 9542

******@********.*** • www.linkedin.com/in/sanjarahmadov • www.github.com/sanjarahmadov E D U C A T I O N

COLUMBIA UNIVERSITY

MS in Operations Research - GPA: 3.54/4.00

• Graduate coursework: Analysis of Algorithms, Statistical Machine Learning, Statistical Inference, Neural Networks and Deep Learning, Optimization, Probability Theory, Business Analytics with R, Applied Integer Programming, Python Based Data Analytics, Stochastic Models, Simulation New York, NY

Sep 2015 – Feb 2017

MIDDLE EAST TECHNICAL UNIVERSITY

BS in Petroleum and Natural Gas Engineering - GPA: 3.65/4.00 (1st Rank in class)

• Undergraduate coursework: Multivariate Calculus, Linear Algebra, Differential Equations, Probability and Statistics for Engineers, Numerical Computing with MATLAB, C Programming Ankara, Turkey

Sep 2009 – Jun 2013

P R O F E S S I O N A L E X P E R I E N C E

Data Scientist: Tata Consultancy Services, Data Science and Advanced Analytics

• Obtaining, cleaning, transforming, and processing data into usable formats (Python, Hive, SQL)

• Building predictive models with range of machine learning, deep learning algorithms (TensorFlow)

• Building and maintaining software tools and packages for core model pipeline of Data Science team

• Organizing and performing complete data analysis and data visualization to communicate findings Troy, MI

Feb 2017 – Present

Data Scientist Intern: Schlumberger, Software Technology

• Predicted oil and gas production using neural networks as classification and regression problems

• Developed neural networks engine where users can input parameters and make predictions (Python)

• Created UI for the engine and integrated it to Windows desktop application (C#, .NET) Houston, TX

May – Aug 2016

Completions Engineer: BP, Wells

• Designed and delivered complex offshore wells with Distributed Temperature Sensor technology

• Processed and analyzed sensor data to evaluate and visualize well performance (MATLAB, SQL) Baku, Azerbaijan

Sep 2013 – Aug 2015

S K I L L S

COMFORTABLE Python, C++, C, R, SQL FAMILIAR Java, C#, JavaScript, MATLAB DATABASES PostgreSQL, MySQL, Oracle 11g WEB Servlets & JSP, HTML & CSS, XML MISCELLANEOUS TensorFlow, Theano, AWS, Hive, Spark, Linux, Git, Vim, Eclipse, Visual Studio, .NET THEORY Data Structures and Algorithms, OO Programming, Machine Learning, Deep Learning, Statistical Tests P R O J E C T S

Image Classification using Convolutional Neural Networks (Python, TensorFlow, AWS EC2)

• Classified 32x32 RGB images in CIFAR-10 dataset into 10 classes with accuracy of 83% using GPU

• Experimented with various CNN architectures, data augmentation and optimization techniques

• Corrupted images in dataset by dropout and then reconstructed heavily corrupted images with CNN Troy, MI

Mar - Apr 2017

System Design for Direct-Broadcast Satellite Operator (Java, Oracle 11g, JavaScript, HTML, CSS)

• Supported team of 10 as Technical Lead in back-end development with Servlets on Tomcat server

• Designed database schema, MVC model and class relationships for the project

• Implemented graphics tool for admin page to generate various types of charts for data visualization Cincinnati, OH

Feb - Mar 2017

Spoken Language Understanding Using LSTM Neural Networks (Python, Theano, AWS EC2)

• Replicated main experiments of Spoken Language Understanding Using LSTM (K. Yao et al., 2014)

• Implemented shallow/deep Recurrent Neural Networks (RNNs) and shallow/deep Long Short-Term Memory neural networks (LSTMs) to assign correct label to a word in a given sentence

• Achieved almost the same results, with 0.5% deviation, as the paper, highest being 95.09% accuracy New York, NY

Oct - Dec 2016

Event Recommender (Python, MySQL)

• Developed web scraper for obtaining and storing event information in NYC

• Created UI where users can create accounts and interact with events through terminal

• Used KNN classification to recommend events based on users’ likes and dislikes New York, NY

Oct - Dec 2015

Predicting Online News Popularity (R)

• Performed feature selections (e.g. L1 regularization) and cross validations to build predictive model

• Predicted unpublished articles’ popularity with 70% accuracy and suggested how to improve it New York, NY

Oct - Dec 2015

Statistical Analysis on Tech Stock Data (R, Shiny) - https://sanjar.shinyapps.io/Project

• Created shiny app to visualize and do statistical analyses on tech stock data from Google Finance

• Features include hypothesis testing, parameter estimation, confidence intervals, linear regression, etc. New York, NY

Sep - Dec 2015

Contact this candidate