Dmitrii
Malofeev

Automated enterprise presentation generator powered by retrieval-augmented LLMs.

• Integrated a RAG pipeline over corporate documents, bridging UI and LLM capabilities.
• Shipped a scalable, production-ready MVP under strict hackathon time constraints.

QR-quest platform for the Student Union — students scan codes around campus via a Telegram MiniApp.

• Researched and hardened the security model so the quest could not be cheated or replayed.
• Engineered the backend to sustain 1,000+ concurrent students without degradation.

Modernizing a 15-year-old enterprise system — turning unreadable legacy into an AI-driven dev tool.

• Built a custom MCP server interfacing with severely outdated, undocumented Russian C# libraries.
• Designed a RAG pipeline ingesting legacy documentation so an AI assistant could autonomously write valid C# macros.

End-to-end mobile MVP digitalizing field engineering services.

• Designed and shipped a mobile-first interface tightly integrated with the 1C Enterprise backend.
• Streamlined work-order and completion-certificate flows, cutting administrative overhead for engineers.
• Defended the technical architecture before the jury — recognized for logical design and market readiness.

AI-powered platform to reduce student information noise and automate content auditing.

• Built a full-stack system with FastAPI + React/TypeScript, deployed via Docker.
• Integrated an LLM service that audits posts and flags non-relevant content.
• Delivered admin dashboards, audit logging, and secure authentication.

Enterprise-grade ChatGPT alternative with dynamic model routing, multi-agent workspaces, and secure code execution.

• Delivered an OpenAI-compatible FastAPI backend integrated with a customized OpenWebUI frontend.
• Built intent-based model routing and multi-agent workspaces with secure E2B sandboxing.

Microservice bridge that mirrors messages between Telegram and Max in real time.

• Designed a distributed system of specialized microservices with async messaging.
• Built a Telegram Mini App to manage Max direct messages inside Telegram.
• Implemented reliable mapping for edits, deletions, and media mirroring.

Distributed microservice platform that personalizes campus news with AI relevance scoring.

• Architected a decoupled system across API, bot, web front, and ML services.
• Orchestrated Docker Compose deployments and managed dependencies via Git submodules.
• Delivered a Telegram bot as the primary client for personalized feeds.

AI task manager that turns incidents and large briefs into actionable tasks in minutes.

• Built the kanban UI, AI draft inbox, and real-time task updates.
• Implemented core API pieces: RBAC, task management, and WebSocket events.
• Delivered a mobile PWA flow plus integrations with webhooks and ML workers.