0 人推荐


在线 离线
I am researcher in wide range of scientific fields. Implemented Machine Learning methods for oil pumping, stock trading, recommender systems, detection anomalies in business' KPIs, object tracking tasks. Frameworks: Matlab, Python (pandas, numpy, sklearn, keras, nltk, scipy, matplotlib). Methods: Deep Fully Connected Neural Networks, Convolutional Neural Networks, Ansamblis of Trees (Boosted Trees, Random Forest), SVM. Have huge experience in feature engineering. Understand how features may be extracted, how to avoid data leakage, how to estimate quality of ML algorithms. I am good at time series processing: Fourier analysis, Wavelets, Correlation functions, etc. PhD of Moscow Institute of Physics and Technology.
$20 USD/hr
  • 100%完工率
  • 100%预算內
  • 100%准时性
  • 25%重雇率





Oct 2017 - Feb 2018 (4 months)

Work as an analyst in the department of data mining. Identification and localization of anomalies in daily and intraday indicators of business performance. Stack of technologies: Matlab, Python (pandas, numpy, sklearn, keras), DSP. • Extracting features from time series • Classification models learning • Correlations between signals


Sep 2016 - May 2017 (8 months)

Developed a system that allows to classify events in the market in real time on events occurring "on their own" and as "reaction". Developed pattern recognition algorithms to identify hidden algorithms for the execution of large orders (Randomizers). Stack of technologies: Matlab, Python (pandas, numpy, sklearn, keras), DSP. • Classification of events on the exchange • Determining of the presence of "big" players. • EM learned graphical models

Search/recommendation engine developer

Aug 2015 - Sep 2015 (1 month)

I developed recommendation/search engine using text data mining, pattern recognition and user-user collaborative filtering algorithms. Stack of technologies: Matlab, Python (pandas, numpy, sklearn, keras)

Leading Software Developer

Jan 2013 - Oct 2017 (4 years)

Mathematical modeling and data analysis: • Processing of telemetry data (readings of a large number of sensors and other pipeline parameters) • Building a decision-making system based on data • Digital signal processing to extract features indicating a non-standard state of the system • Building learning algorithms (NN, autoencoders, ansamblis) to identify and localize potentially dangerous states Stack of technologies: Matlab, Python (pandas, numpy, sklearn, keras)


Sep 2011 - Jan 2013 (1 year)

Development of automated trading algorithms on the futures market. Development of trading robots and their implementation with C #. Analysis of of large volumes of data, building statistical models, self-learning algorithms.

QA Engineer

Jan 2011 - Jul 2011 (6 months)

Testing of mobile applications, bugs management, preparation of test cases.



2007 - 2015 (8 years)


Machine Learning (2016)

Stanford University

Neural Networks for Machine Learning (2017)

University of Toronto

Introduction to Data Science in Python (2017)

University of Michigan

Digital Signal Processing (2014)

École Polytechnique Fédérale de Lausanne


Application of digital signal processing algorithms to identify the unsteady point source of signal

Detailed information: [login to view URL]