Doston Kamilov

Built and maintained 8+ NestJS microservices on a telemedicine platform (appointments, video consultations, billing and doctor–clinic workflows) using JSON-RPC and message brokers. Led migration from RabbitMQ to Apache Kafka, delivering it in the billing service and driving rollout to doctor–clinic interaction, including a transactional outbox for reliable publishing, idempotent consumers, dead-letter topics, and shared event contracts. Deployed LiveKit (WebRTC SFU) and built the NestJS video service around room lifecycle, access-token issuance and participant management. Integrated Centrifugo for real-time appointment status updates. Worked with third-party face recognition and reduced costs ~30% via a free pre-flight consent check, retry grace window, and capped verification attempts. Added Prometheus metrics and Grafana dashboards for throughput/latency/error-rate and third-party 4xx/5xx monitoring. Initiated agentic development with Claude Code and MCP integrations; used Redis

Sole backend engineer on a restaurant automation platform. Independently designed system architecture and delivered a working prototype. Built an on-premises backend in Go to run directly on an Android master tablet acting as the local server for client tablets across a venue; implemented authentication and a fine-grained role/permission system modeled on Django, plus thermal receipt-printer integration and separation of responsibilities between device roles. Used SQLite on-device and Server-Sent Events for server-to-client push to keep the prototype simple for unreliable connectivity. Designed the NestJS cloud layer for multi-site synchronization and remote management.

Built a NestJS/Prisma/PostgreSQL backend for an AI-assisted veterinary diagnostics system delivered with Samarkand Veterinary University. Implemented registration of animals and veterinarians across districts, electronic animal health records, and prediction of 100+ diseases from 30+ clinical indicators. Created a veterinarian feedback loop to feed practitioner corrections back into model retraining. Owned CI/CD end to end with GitHub Actions publishing to GitHub Container Registry and containerized frontend/backend/ML model for automated server deployment. Also built the frontend in Next.js/React with Tailwind, Zustand, shadcn/ui and Feature-Sliced Design. Taught Go backend development and software engineering foundations, running code reviews, Q&A sessions, and individual technical guidance for students.