Backend systems designed for scale, clarity, and change.

Lead the architecture and delivery of a hotel booking service for Russ while modernizing other company platforms.

• Designed the backend of a high-load distributed platform with six greenfield Python services and more than ten external integrations.
• Modeled more than 30 business entities and implemented asynchronous workflows across PostgreSQL, MongoDB, Yandex Object Storage, Redis, and RabbitMQ.
• Introduced a CQRS-based hybrid storage architecture that improved content delivery performance by 45% while preserving consistency for critical data.
• Migrated 200,000 users from a legacy database to a modern JWT-based authorization flow and isolated FastAPI and Django migration domains, cutting pilot delivery time in half.
• Established per-service Docker Compose environments, CI/CD pipelines, 400+ tests, code review, onboarding practices, individual development plans, and architectural decision records.

Developed the backend and frontend of a virtual tabletop platform for running online tabletop role-playing games.

• Designed an event-driven architecture with Apache Kafka to synchronize chat, dice rolls, and token movement across microservices.
• Built a character persistence module with dynamic schemas using MongoDB and PostgreSQL as a hybrid data model.
• Designed a FastAPI and WebSocket layer that synchronized player actions with latency below 50 milliseconds.

Owned a web product that recognized, annotated, and pre-filled entities from clinical documents to automate the work of clinical analysts.

• Added request queuing and caching, persisted processing results in PostgreSQL, and implemented reliable delivery to the client application.
• Introduced Prometheus metrics and Grafana dashboards for product and operational visibility.
• Maintained and scaled an API used by at least five external projects across other teams.
• Integrated Claude 3.5 Sonnet and increased the number of supported UI forms fivefold.
• Delivered more than 100 unit, integration, and end-to-end test scenarios while modernizing CI/CD and coordinating requirements and sprints with customers and engineering teams.

Developed and maintained software for testing computer vision algorithms and researching how image processing affects neural network performance.

• Built an internal application for testing neural network-based computer vision models.
• Deployed the CVAT annotation service in the organization's internal environment.
• Validated and corrected more than 20 neural network models and their datasets.
• Researched the impact of image processing algorithms on model quality.