— Full ML model development lifecycle: requirements gathering, research, experiments, production deployment
— Applying LLM for automated data labeling, extraction, and classification
— Developing pricing, product matching, categorization, and forecasting algorithms
— Building data processing pipelines and data quality systems
— Using automated testing to ensure the quality of data and models

— Conducted R&D and deployed 8 ML models to production
— Designed an AI system based on LLM and RAG to maintain the product type catalog, reducing the need for manual labeling
— Implemented ML model for product matching, reducing unmatched supplier offers by 2x
— Developed duplicate scorer, reducing implicit product duplicates by 60% with 95% precision
— Developed price regressor with confidence intervals, simplifying supplier offer moderation
— Implemented price anomaly detection, identifying unit-of-measure and data entry errors
— Implemented last-mile median delivery time forecasting with less than 15% error
— Optimized data processing pipelines, reducing memory consumption by 5x

— Refining existing services
— Implementing machine learning techniques
— Generating ad-hoc reports
— Investigating data consistency issues
— Researching solutions for current business needs

— Created a Rate Calculator for container transportation (achieving 11% RMSPE), featuring an intuitive interface, performance monitoring, and regular retraining capabilities
— Built BI dashboards for container transportation, enhancing operational decision-making and reducing manual reporting efforts
— Developed an OCR service with 95% precision in document validation, significantly reducing dispatcher workload
— Implemented an Email AI system using ChatGPT, automating email handling and reducing response times

— Implementing machine learning techniques
— Topic text classification
— Searching for significant features, which cause product dissatisfaction
— Helping SEO-specialists to construct effective promotion strategies

— Development of price recommendation model from data analysis to preparing for deploy
— Developed a backlink search system, optimizing search speed to 3 seconds and improving SEO relevance and customer satisfaction

— Choosing the technical development strategy and tools, suggestion/discussion of implementation ways. Code review
— Creating tools to analyze and manage ads
— Synchronization of statistics, state of moderation
— Implementing of new formats. Error handling
— Deployment of the system. Backups. System state monitoring

— Modifying the API in such a way it can be used for external users and the frontend
— Synchronization and display of advertising performance from Yandex.Metrica data
— Adding a section for problem detection in the system
— Implementation of interactive widgets: weekly schedule, bulk copying, ad preview
— Frontend migration from CoffeeScript and jQuery to ES6 and React
— Correction of the layout and use of the single style within the entire user account

— Real estate project support and development http://o-din.ru/
— Beauty specialists qualification project support and development http://e-academie.ru/

— Graduation work rewriting https://github.com/resivalex/student-progress-on-rails
— Trampoline center website development https://github.com/resivalex/zaskok

— Website development http://pskovskie.ru/ from scratch
— Development of a multi-component network application in a team of five members using the Scrum methodology
— Full implementation of an important component interacting with the exchange
— Unit-testing with Catch framework
— Partially designing and implementing a client application on Qt
— Using design patterns
— Code review

— Improvements of the trading engine
— Exchange log parsing and analysis
— Features implementing for a transaction logging software
— Development of a program for emulation an exchange
— Sending email to users of specialized sites
— Automatic exchange log downloading
— Using system of version control