2026
5 years
Professional experience
Backend & Software Engineer focused on backend systems, architecture, and infrastructure.
Working across the full system lifecycle: from initial design decisions to deployment and production support.
About
Backend & Software Engineer with 5 years of commercial experience in developing software systems. I specialize in building and evolving complex backend solutions, contributing to system architecture design, product development, and technical decision-making throughout the entire software development lifecycle.
I am also involved in research activities in the field of distributed systems and load balancing algorithms. Winner of LCT 2025 and prize winner of LCT 2024.
I value openness, independence, and continuous growth. In both work and life, I prefer to deeply understand things and consistently bring what I start to completion.
Professional feedback
Employer recommendation
During his time with the company, Grigory proved to be a responsible and independent professional. He understands technical requirements, proposes practical implementation options, and sees tasks through to a working result. He demonstrates a systematic approach and attention to detail, and is willing to learn new tools and apply the knowledge he gains.
Middle Backend & Software Engineer
Employment: Aug 2024 — Sep 2026
Skills
Experience
I develop backend services and integrations between ESZK (Unified Customer Care Service) and Sber’s internal systems. I design microservices, REST APIs, and event-driven integrations based on Apache Kafka, and contribute to technical solution design and integration contracts together with analysts and cross-functional teams. Development is carried out in Python 3.13 and Java 21 using Spring MVC, Spring Kafka, FastAPI, Pydantic, and SQLAlchemy. Services are deployed to Kubernetes (Platform V DropApp) and go through the full DEV -> PSI -> PROD delivery pipeline using Jenkins and ArgoCD.
I develop integration infrastructure and data exchange processes across internal banking systems, including Greenplum- and Kafka-based workflows. I standardized approaches to system integrations, reducing the time required to connect new integrations by 23%; introduced additional DTO validation and contract controls, reducing integration errors by 75%; and optimized Kafka event processing, reducing processing latency by 31%. I also contribute to CI/CD pipelines, Kubernetes infrastructure, Istio-based Service Mesh configuration, and secure secrets management through SecMan/Vault.
Python 3.13 · Java 21 · Spring MVC · Spring Kafka · Apache Kafka · Confluent Kafka · aiokafka · Pydantic · SQLAlchemy · PostgreSQL · Greenplum · FastAPI · Kubernetes · Platform V DropApp · Istio · Jenkins · ArgoCD · SecMan/Vault · Maven · Bitbucket
Developed and maintained a software platform for advertising delivery and a distributed network of IoT devices. Worked across the full development lifecycle — from system architecture, API design, and data modeling to backend and frontend implementation, testing, CI/CD, and production infrastructure support. Applied DDD, Clean Architecture, and a microservice-oriented approach to system design.
Designed a centralized IoT device management infrastructure based on OpenBalena and migrated a legacy advertising delivery algorithm from Ruby to TypeScript, which, together with infrastructure improvements, helped increase streaming availability (uptime) from 80% to 98%. Automated server deployment with Ansible, introduced Zabbix monitoring and a corporate mesh VPN, and developed internal products including an advertising campaign pricing system and a Telegram bot for partner communication and automated reporting.
Python · FastAPI · SQLAlchemy · PostgreSQL · Redis · TypeScript · React · Next.js · Docker · Nginx · Ansible · OpenBalena · Zabbix · DDD · FSD · Clean Architecture
Developed web interfaces and the frontend of an MVP using JavaScript, jQuery, and Bitrix CMS, integrating client applications with backend APIs, payment systems, and external services. Contributed to frontend architecture, performance optimization, code review, testing, and the ongoing development of the existing codebase.
Standardized styling across existing templates and introduced a UI Kit and reusable JavaScript components. Established separate development and production environments for Bitrix CMS and introduced a Git-based development and deployment workflow.
JavaScript · jQuery · Bitrix CMS · PHP · HTML · CSS · Nginx · Git
Developed and maintained frontend applications based on Bitrix CMS with integrations to PHP backend APIs. Worked on extending the existing codebase, refactoring, SEO, page-load performance optimization, and testing.
Separated development and production environments by introducing an isolated testing environment and established a complete Git workflow with development branches, pull requests, code review, and controlled production deployments.
JavaScript · jQuery · Bitrix CMS · PHP · HTML · CSS · Nginx · Git
Yantar+
Developed backend services in C#/.NET and ASP.NET Core for processing and synchronizing data with an Android application. Designed REST APIs and MySQL and SQLite databases, working with Entity Framework Core, migrations, indexes, SQL queries, asynchronous processing, and multithreading, while also maintaining the client-side Android application in Java.
Redesigned the database update pipeline by eliminating an intermediate SQLite-to-DBF conversion, reducing update time from 35 minutes to 3 minutes and fully automating the data-loading process. Also implemented asynchronous processing and automatic data synchronization between backend services and the Android application.
C# · .NET · ASP.NET Core · Entity Framework Core · REST API · MySQL · SQLite · SQL · Java · Android · Git
Education
Specialization: Applied Mathematics.
Educational track: “Algorithms for Science-Intensive Software.”
During my studies, I focused on mathematical and algorithmic methods, data structures, computational approaches, and software design.
Specialization: Applied Informatics.
Educational track: «Artificial Intelligence and Machine Learning».
The program combines advanced computer science studies with practical applications in machine learning, NLP, computer vision, robotics, MLOps, and data processing systems.
Professional retraining program at NUST MISIS: «Design and Development of Network Applications».
The program covered client-server application development, Astra Linux administration fundamentals, and MS SQL database administration. The total workload was 250 academic hours, with the final assessment completed with an “Excellent” grade.
Sber's corporate training program focused on developing engineering expertise, exploring modern software development practices, and working with enterprise IT infrastructure.
The program covers Agile and Sbergile methodologies, Spec-Driven Development, and the fundamentals of containerization and application orchestration with Kubernetes. Particular emphasis is placed on distributed data processing and storage systems using Greenplum, covering both foundational and advanced concepts of analytical database management systems.
Android Application Programming and Development.
During the program, I gained hands-on experience in mobile development and built an individual Android project. I completed the program with honors and received a Certificate with Honors.
Projects
NTV Club — a loyalty program web platform with a Telegram bot, allowing users to register and authenticate via email, phone number, and OAuth, browse personalized offers, add promotions to favorites, and activate them. Content and business entities are managed through Strapi Headless CMS.
Fully developed the system from scratch and was responsible for the entire development lifecycle: designing the architecture, building the frontend with Next.js/React and backend services with FastAPI, configuring PostgreSQL using SQLModel, integrating with Strapi CMS, and developing a Telegram bot with Aiogram. Implemented a dedicated authentication service with OAuth and JWT sessions, user workflows for promotions and accounts, as well as deployment and monitoring of the production environment.
Python · TypeScript · Strapi Headless CMS · Aiogram · SQLModel · FastAPI · PostgreSQL · Docker · React · NextJS · SASS
IoT Infrastructure Modernization — a project focused on improving streaming reliability and enabling centralized management of IoT devices.
Designed and implemented a centralized IoT device management platform based on OpenBalena, providing automated application deployment, BalenaOS updates, device monitoring, and remote administration. Developed the architecture for component interaction and prepared the infrastructure for scalable device fleet management.
Automated server provisioning and configuration using Ansible and Bash, configured secure VPN/OpenSSH access to IoT devices, and set up a containerized environment. As a result, improved system reliability and increased streaming availability (uptime) to 99%, while making device management centralized and more reliable.
OpenBalena · OpenVPN · Ansible · Let's Encrypt · Bash · Git · Docker · balena-cli · OpenSSH
An internal fullstack platform for managing advertising campaigns, commercial proposals, and advertising placements. The system automates advertising campaign cost calculations and key advertising metrics, enables management of placement locations and packages, and supports creating, approving, publishing, and tracking commercial proposals throughout their lifecycle.
Independently developed the system from scratch and took ownership of the full development lifecycle: designing the architecture, implementing backend APIs and business logic with FastAPI and PostgreSQL, developing the frontend with React and TypeScript, and delivering key user workflows. Implemented role-based JWT authentication, advertising campaign cost calculation, calculation snapshots storage, Excel data export, and advertising placement management. Developed frontend interfaces with server-state caching, filtering, forms, and an interactive Leaflet map for selecting and visualizing advertising locations. Configured deployment, CI/CD pipelines, and production environment monitoring.
React · TypeScript · Material UI · Python · FastAPI · Alembic · Poetry · SQLAlchemy · PostgreSQL · Docker · Vite · Leaflet · Pydantic
A centralized monitoring system for the Novikov TV server infrastructure, designed to track Linux server health, resource consumption, and proactively detect infrastructure issues. The system monitors 14 servers, automatically collects CPU, RAM, disk space, and filesystem metrics, generates dashboards, and sends notifications when triggers are activated.
Fully configured the Zabbix 7.4 monitoring infrastructure, automated Zabbix Agent 2 installation and removal using Ansible, and implemented secure server auto-registration via PSK authentication. Organized host grouping, configured email alerts, and documented deployment and maintenance processes. Using the implemented monitoring system, identified issues with inode exhaustion, high RAM consumption, and recurring server restarts, then optimized resource usage across production servers.
Zabbix 7.4 · Zabbix Agent 2 · Docker · Docker Compose · Bash · Linux · PSK/TLS · SMTP · Yandex Mail
A fullstack prototype of a medical system for fetal condition monitoring and cardiotocography (CTG) analysis, developed for the Leaders of Digital Transformation 2025 hackathon, where our team became a winner. The application displays fetal heart rate and uterine activity in real time, calculates and visualizes clinical metrics, FIGO status, and alerts, and provides STV and hypoxia risk predictions. It also supports examination history, including trend analysis, detailed review of individual CTG records, and visual comparison of two examinations.
Developed both the client and server sides of the system, including real-time data streaming via WebSocket, interactive medical dashboards with React and TypeScript, state management with Redux Toolkit, and API data validation with Zod. The backend was built with FastAPI, asynchronous SQLAlchemy, and dependency injection via Dishka; the ML module processes CTG time-series data, calculates STV, accelerations, and decelerations, generates predictions, and analyzes examination trends. The system also includes LLM-based anamnesis summarization and containerized deployment of backend components.
Python · Dishka · scikit-learn · CatBoost · Pandas · FastAPI · SQLAlchemy · WebSocket · Alembic · TypeScript · React · D3 · React Redux
A client application for Raspberry Pi designed for automated playback of advertising and informational video content on digital displays. The application synchronizes with the server, retrieves up-to-date playlists, downloads and stores required media files locally, and automatically updates the content being played. It also supports offline operation during temporary server outages using locally stored data and media files.
As part of the project, I developed and maintained the client-side application using TypeScript/Node.js. I implemented playlist and media synchronization with the backend API, local storage management, and video playback via OMXPlayer. I also implemented playlist update handling, media file integrity checks, playback statistics reporting with local buffering, and additional informational content display. The application runs on Raspberry Pi in a Docker environment and uses systemd for service management.
Raspberry Pi · systemd · Linux · BalenaOS · BalenaEngine (Docker) · TypeScript · OpenBalena · OMXPlayer
A secure proxy infrastructure built from scratch to provide employees with reliable access to external resources. The system is deployed across 5 servers in multiple countries and provides centralized access to proxy services for 45 employees, supporting multiple connection protocols, split routing, load balancing between nodes, and centralized client configuration management.
Designed and implemented the infrastructure from scratch: automated server provisioning with Ansible, configured 3x-ui, Xray, TLS/REALITY connections, and traffic routing across multiple nodes. Implemented split routing to direct different categories of traffic through dedicated servers, as well as load balancing between available nodes. Optimized the Linux network stack using `sysctl`, including BBR configuration, TCP parameters, connection queues, and basic network hardening. Configured server monitoring through Zabbix with active metric collection. As a result, reduced the deployment time of a new proxy node from several hours to 10–15 minutes.
3x-ui · Xray-core · VLESS · REALITY · XHTTP · Trojan · Hysteria2 · Ansible · Linux · TLS · BBR · Load Balancing · Split Routing
A fullstack system for procurement planning and analytics, developed for the Leaders of Digital Transformation 2024 hackathon, where our team won 3rd place. The solution supports procurement and product management, purchase history and inventory analysis, similar-product search, and demand forecasting for monthly, quarterly, and yearly periods. For regularly purchased products, the system can predict the next procurement, including expected dates, volume, and cost.
Developed components of the fullstack solution, including a FastAPI backend with PostgreSQL, Redis, and JWT-based authentication, integration with a separate ML service, and web interfaces built with React and Redux Toolkit. The ML component used RuBERT and vector similarity for product matching, time-series forecasting models, and Whisper for voice-query recognition. Also developed an Aiogram-based Telegram bot for creating and managing procurements, analyzing products, viewing statistics, and running forecasts directly from the messenger.
Python · FastAPI · SQLAlchemy · PostgreSQL · Redis · JWT · Pydantic · Pandas · NumPy · scikit-learn · PyTorch · Transformers · RuBERT · ARIMA · Whisper · Aiogram · React · React Redux · JavaScript · Docker · Caddy
A fullstack Telegram Mini App featuring a digital Tibetan calendar with detailed information about specific days, lunar phases, events, and recommendations. The application supports weekly, monthly, and yearly calendar views, user accounts, onboarding, and a paid subscription model, as well as synchronization of calendar events with Google Calendar.
Developed the backend and business logic with FastAPI using domain, application, infrastructure, and presentation layers. Implemented Telegram Web App authentication with JWT, user and subscription management, trial periods, promo codes, Robokassa payment integration, and scheduled background jobs for subscription monitoring. Built the React/TypeScript frontend with Redux Toolkit, routing, Zod-based API validation, and FSD architecture, as well as a separate Aiogram-based Telegram bot for user registration and launching the Mini App.
Python · FastAPI · SQLModel · Alembic · PostgreSQL · AsyncPG · APScheduler · JWT · Google Calendar API · Google OAuth 2.0 · Robokassa · Aiogram · TypeScript · React · React Redux · Zod · CSS · FSD
A computer vision system for automatically detecting polar bears in large-scale 8K aerial images of the Arctic. Developed as a team project as part of an image processing course at NUST MISIS. The application analyzes an image, determines whether a polar bear is present, and, when detected, localizes it in the original image. The algorithm is based on classical computer vision techniques and does not require training an ML model.
Developed the core Detector class and the image processing pipeline in C++ and OpenCV, including HSV conversion, adaptive filtering based on image brightness, color segmentation, binary mask generation, and detection of the region with the highest density of candidate pixels. Also implemented detector evaluation using Precision, Recall, F1, and IoU metrics, as well as parameter tuning with parallel processing via OpenMP. On the 8K test dataset, the algorithm achieved 80% Precision and 76.2% Recall, with an average processing time of approximately 1 second per image.
Python · C++ · OpenCV 4.9 · OpenMP · CMake · vcpkg · Computer Vision · Image Processing
Contact