AI engineer building production AI products: LLM agents, RAG with measured quality, MCP tools.
I take a product from data model to deployed, monitored service, and I make AI output reliable:
validated, testable, traceable to source data.
Before software I spent 10+ years running production infrastructure (Linux, Windows Server, Azure,
telecom billing, networks). So the systems I build are deployed, secured and monitored, not just written.
📍 Ukraine · remote · griban.dev · LinkedIn
HealBot.pro — AI front-desk assistant for private clinics.
Talks to patients in Telegram, Viber, WhatsApp, Messenger, Instagram and a website chat, collects complaints,
routes each patient to the right specialist and books the appointment. Docker, automated deploys, payments, bot protection.
(private repository)
Cryonic CRM — AI-first CRM for outbound sales. Next.js + NestJS, self-hosted Supabase (PostgreSQL + pgvector), background AI jobs and call transcription (Whisper) in a Redis queue, subscription billing, versioned public API. (private repository)
| Project | What it is |
|---|---|
| memlab | Project memory for coding agents: graph-aware code retrieval + decision memory, MCP server, pre-registered benchmarks against other tools. Python. |
| dokploy-mcp | MCP server exposing the full Dokploy API through 4 gateway tools (~1% of context) or 500+ generated tools, with safety annotations. TypeScript. |
| myhealth-mcp | Self-hosted, FHIR-native MCP server for citizen-controlled health data in the EU. |
| agents-md-in-the-wild | 480 real agent instruction files (AGENTS.md, CLAUDE.md…) from 272 top repositories, refreshed weekly, with patterns measured. Python. |
| talersme | AI assistant that helps European SMEs integrate GNU Taler payments. |
| N8N-AI-Starter-Kit-Advanced | Production n8n + AI services deployment: monitoring, CI/CD, disaster recovery. |
AI: Claude & OpenAI APIs, agents & tool calling, MCP, RAG, embeddings, pgvector, evals, structured outputs
Backend: Python (FastAPI, Pydantic), TypeScript (Node.js, NestJS), PostgreSQL, Supabase, Redis, queues
Frontend: Next.js, React
Ops: Docker, nginx, Dokploy, Railway, CI/CD, Prometheus, Grafana, Sentry, Linux, Azure
Automation: n8n, Make, Zapier
Open to freelance and contract work: AI assistants and chatbots, RAG over documents, agent systems, AI features in SaaS products. I communicate in writing (English, Ukrainian, Russian).