AI engineer. I build and run production systems: Python backends, LLM integrations, Telegram services and automation pipelines — from first commit to live deployment on my own servers.
I work AI-first (Claude Code, Cursor) and own the parts that tooling can't: architecture, review, acceptance and operations — monitoring, watchdogs, incident triage.
- Backends & APIs — FastAPI, PostgreSQL/PostGIS, SQLite, Celery, Redis, Alembic
- LLM integrations — OpenAI / Anthropic APIs, RAG, prompt engineering, n8n workflows
- Telegram services — aiogram bots, payments (Telegram Stars), schedulers, alerting
- Infrastructure — Linux, systemd, Docker, Nginx, Let's Encrypt, Cloudflare, VPS deployment
- Data & ML — pandas, scikit-learn, CatBoost, walk-forward validation, backtesting
| Project | What it is |
|---|---|
| ai-ticket-triage | Webhook → Claude structured-output triage → SQLite → n8n automation. FastAPI reference implementation |
| task-marketplace-api | Hyperlocal gig marketplace REST API — FastAPI, PostgreSQL/PostGIS, Celery, Redis, MinIO, Docker |
| binance-futures-backtest | Walk-forward backtesting framework, 3 years of tick data, explicit no-lookahead guarantees |
| algorithmic-trading-demo | Reference architecture for algorithmic trading systems |
| tennis-ml-models | CatBoost models for ATP/WTA match outcomes with calibration reports |
Several current projects are commercial and closed source — Telegram service with a live billing loop, a 24/7 market data collector on VPS, and a generative media pipeline. Happy to walk through architecture and demo them on a call.
Telegram: @gotodata · Email: multiakall@gmail.com