Yuriy Osychenko Email & Phone Number

Moscow, Moscow City, Russia

Moscow, Moscow City, Russia

• Developed System Design for offline retail recommender system
• Led and owned the whole process: from drawing initial architecture, to generating requirements, writing documentation and communication with members from other teams (backend and devops) and product owners
• Implemented one of the services from mentioned architecture, namely Recommender Models Execution Service (python3, FastAPI, docker)
• Helped Data Science team bring their recommender models and algorithms from Jupyter notebooks into production
• Developed internal framework and project structure templates for easy conversion from Jupyter notebooks to proper Python application (FastAPI, pytest, poetry, nox)
• Built CI/CD pipelines from the ground up in order to automatize ML models deployment (docker, k8s, jenkins)

Moscow

• Developed model for quality assurance of STT generated dialogues
• Model was trained on augmented data from ODS SlackDeveloped model for assessing customer satisfaction
• Basic SVM+TFIDF model trained on client phrases (scikit-learn, eli5)
• Used eli5 library in order to interpret model decisions
• Model has helped to emphasize key issues in costumer serviceProposed method for checking mistakes in data annotation
• Method utilizes already trained model. It compares confidence of predicted classand annotated class. If model is not confident in annotated class, it is probablyannotated incorrectly.

Remote

• Created German grammar correctness analyzer application based on static rules
• Created Transformer based model in order to solve NER and Aspect-Based Sentiment Analysis problems using restaurant reviews data
• Analyzed restaurant reviews data in order to gather insights about business daily operation: whether or not people liked the food; whether the service was fast and pleasant, or there were rude and slow waiters; was it.

Volgograd, Russia

• Integration project for TOP-3 Credit Bureau in Russia
• Developed parallelized ETL PL/SQL procedures to move 100mil+ rows of data for Oracle 12c database.
• Developed analytical SQL reports.Migration project from Oracle to PostgreSQL database
• Migrated PL/SQL code to PL/pgSQL code.Information portal for a client in a Government Regulatory sector
• Created Talend jobs to gather sparse data from different sources, e.g. databases, SFTP-servers, flat files etc.
• Developed Pentaho ETL jobs for internal projects.

Master'S Degree, Software Engineering

Thesis: "Method for recognizing human emotions based on EEG data"

Bachelor'S Degree, Software Engineering