Uladzislau Sataneuski

Built systems for social media text analysis using Python. Most exciting for me were machine learning systems for user location detection using explicit location references and implicit “local” tokens, sentiment analysis model for very slangy lexis with lack of annotated data. Also built different crawlers and data processing pipelines using Spark, Kafka, MongoDB, Postgres, Redis. Was also responsible for finding bottlenecks and speeding them up using better algorithms or C++ extensions.

Built experimental Google Analytics 360 feature recommender using Python. It allows consultants generate good ideas of what features could be useful for their premium clients. Successfully presented results to product managers.

Developed internal machine learning project using Python. Implemented complicated algorithm for data clustering including machine learning model as its part. It made data analysis possible, because non-aggregated data items contain much less meaning. Improved algorithm computational performance from approximately few years to one night without affecting results quality using advanced algorithms and features of problem and data. Analyzed resulting data using statistics and machine learning methods.

Implemented module for simulating influence of ultrasound scanner noise on the resulting image using C++. This module was a part of ultrasound scanner simulator made by group of students to provide ability of running simulations using ultrasound scanner.

Big Data Processing

Data Science, Algorithms and Data Structures, Programming

Discrete Mathematics and Algorithmics