ML engineer focused on building production systems for algorithmic trading and sports analytics.
All projects are validated on real out-of-sample data. Working strategies are closed source.
| Project | Stack | Description |
|---|---|---|
| predicttrack | Python · Flask · Telegram · Claude API | SaaS tracker for Polymarket + Kalshi traders |
| algorithmic-trading-demo | Python · CCXT · pytest | Modular architecture for algo trading systems |
| tennis-ml-models | CatBoost · scikit-learn | Research: ATP/WTA prediction — AUC 0.75, no exploitable edge found |
| 1c_petconfig | 1C:Enterprise | Custom configuration built from scratch |
ML: CatBoost · XGBoost · LightGBM · SHAP · walk-forward validation · bootstrap CI
Trading: statistical arbitrage · delta-neutral strategies · Binance/Bybit API
Backend: Python · Flask · SQLite · Telegram Bot · APScheduler
Data: pandas · DuckDB · Parquet · Playwright scraping
1C:Enterprise development