Responsible for the creation of test-database and manipulations with it. It was used Cats, Doobie and H2 to complete this task. Worked with the services of the project, optimized a little bit of validation part of this service (constraints). Some tasks were done based on Akka for concurrent processing of user data and requests. Another responsibility was to cover the project with a lot of tests.
A platform for planning and organizing international traveling. Includes an advanced system for booking plane tickets and a variety of ground and water transport, building travel routes, an advanced search system, customer care, email distribution, and others.
Technologies Stack:PostgreSQL Scala Akka REST Swagger
Travel & Tourism
Recommendation system for a book shop
Responsible for a Recommendation System micro-service, writing tests and bugfixes. Developed Recommendation Engine based on Spark Structured API (ALSModel), Streaming API (implemented the BSGD algorithm for it) and Akka. Created a lot of Actors for Spark model synchronization and updating without any downtime. Development of REST controllers based on a Play Framework later connected to Akka through a level of service classes.
A system that’s used for predicting user’s rating and opinions about books. The system makes predictions based on the opinion of every user. In addition, it calculates user’s and book’s feature vectors alone, so developer/administrator can control it.
Technologies Stack:Cassandra Java Scala Spark Play Framework PhantomJS
Books & Catalogs
Responsible for creation of ML module, development of unit tests and bugfixes. Data engineering (visualization, preprocessing), feature engineering (extraction, generation) and model creation (choosing, validation, testing). Responsible for Spark Estimators and Transformers. Participated in the development of a search engine used for data extraction, in particular, developed highlighter based on Lucene’s base highlighters. All tests are done using the spec2 library.
The program provides the possibility to check how different Data Processors implement GDPR. It splits the document into parts using key-words, saves them in the DB. Another part analyses parts and makes a prediction about the quality of implementation.
Technologies Stack:Java Scala Spark Lucene Hadoop
Team chat for company
Responsible for the creation of some parts of a DAO system of this application, also participated in the development of Spring MVC controllers and service/model layer. Managed configuration files for a Spring IoC. It removed all repeatable code from the application using Spring AOP.
SF-based startup company
For registered users portal provides the ability of exchange messages, files and saving all messages history. A user can point favorites messages, create rooms notifications, create public and private chat rooms, search messages in history archive with the special date or on some text pattern, etc.
Information services & Technologies
V. N. Karazin Kharkiv National University