mafisto

Developed 1 service for routing incoming loan applications to required banks and maximizing profit for applications. Designed and implemented from 0 an end-to-end machine learning system that accurately predicts customer demand with more than 70% accuracy. Clarification of the requirements. Prepared system design and solution to predict top 5 most profitable entities. Boosted the client's revenue. Combined data from 3 different independent sources into 1 dataset, transformed existing features and created new ones, prepared dataset for ML model. Set up train/test ML pipelines. Produced 1 docker image that contains REST service to work with trained ML model. Speeded up ML model inference speed 10x and reduced service response time to 100ms.

Managed the team of 7 people: 5 engineers, 1 tester and 1 business analyst. Onboarded 2 new team members. Code review. Mentored and trained 3 team members to develop their skills and career paths. Deployed and orchestrated 4 microservices across development, testing, and production environments, ensuring seamless integration with existing services; resolved critical infrastructure issues, minimizing downtime and improving system reliability. Developed an interface for engaging with the government system gosuslugi.ru. Built 3 microservices for logging, billing, monitoring, proxy and routing requests between internal bank services. Maintained Gravitee APIM system, created 3 plugins for handling and transforming requests. Writing documentation, planning future tasks and project goals, creating and describing tasks for developers, communication with internal departments. Migrated existing microservice from Java 8 to Java 17 and from Gradle to Maven. Integrated microservices with Bank auth service by using JWT.

Integrated 2 logging services with existing search services created by search department in European Patent Office. Created dashboards in kibana to display requests information. Fixed bugs in existing elasticsearch plugins.

Developed an object detection system for automating the food ordering process in cafes, achieving a 90% accuracy in detecting objects in tray images with food. Optimized ML microservice performance by 3 times. Investigation current ML/DL algorithms for object detection and recognition. Create ML pipelines for test/train, artificial dataset generator, test/train dataset. Suggested new approach for training existing model and adding new classes without training whole model that helps increase speed. Integrate ML models with existing services.

Developed a web application for intellectual property management with advanced free text search capabilities. Developed and designed architecture for new features for web application. Maintained whole developing feature flow: analysis, design, creating back-end part, front-end part, integration and testing. RnD for new features and providing solutions for integration with existing features. Developed and implemented a docker-based online documents editor with seamless integration into the UI. Created user-friendly mail service with advanced editing, formatting, and email sending capabilities. Created new web-components for application. Investigated and analyzed existing plugins/solutions that can be integrated into the application. Migrated project from Vaadin 8 version to Vaadin 11. Created IOS mobile app by using cordova framework. Fixed Vaadin framework memory leak issue and increased application response 5x times. Integrated SSO solution for authorization in application.

Optimized speed in and out zooming map in web application. Increased map response in 2x times.

Install and configure software and hardware. Managed and coordinated 50 computers and 4 servers stationed in Belarus. Monitor performance and maintain systems. Troubleshoot issues and outages. Set up infrastructure (hardware, network etc.) for new offices. Installed Closed Circuit Television.