Vlad Y.

Backend Engineer

Mar 2026 - Aug 2026 • 5 mos

Led the end-to-end development and deployment of a production-ready B2B e-commerce platform integrated directly into Telegram Mini Apps. Handled everything from architectural design to cloud infrastructure and secure payment gateways. Key Responsibilities & Achievements: - System Architecture: Architected the backend utilizing Domain-Driven Design principles (Use Cases & Repositories). Strictly decoupled business logic from frameworks to ensure high maintainability and testability. - FinTech Integration: Engineered secure, asynchronous Webhook handlers for global cryptocurrency gateways (NOWPayments, CryptoBot), establishing a fully automated payment flow with zero chargebacks. - Background Processing: Configured and managed asynchronous task queues using Celery and Redis to optimize system performance and handle heavy background operations. - DevOps & Deployment: Fully containerized the application stack using Docker and Docker Compose, ensuring seamless and isolated deployments across cloud environments. - CI/CD & QA Automation: Set up continuous integration pipelines via GitHub Actions, automating static code analysis (Ruff, Mypy) and unit testing (Pytest) to maintain strict code quality standards.

AI Integration Developer

Jan 2026 - Feb 2026 • 1 mo

Engineered and open-sourced a highly optimized, multimodal AI Telegram bot. Managed the entire development lifecycle—from asynchronous architectural design to hardware-level memory optimization—creating an AI companion capable of processing text, voice, and images in real-time. Key Responsibilities & Achievements: - AI & ML Integration: Successfully integrated low-latency LLM and Vision capabilities via Groq API and OpenRouter. Configured and deployed local zero-shot Voice Cloning utilizing Coqui XTTS. - Asynchronous Engineering: Built a robust, non-blocking backend infrastructure using Aiogram 3 to continuously handle concurrent and diverse user inputs (audio, images, text) without degrading performance. - Hardware Profiling & Optimization: Profiled application memory and optimized resource allocation to run heavy machine learning models efficiently on constrained hardware (15GB T4 GPU). - Performance Stability: Successfully implemented advanced memory management techniques to prevent Out-Of-Memory (OOM) errors, ensuring continuous and stable system execution under load.