MBATS
Autonomous Trading System
Overview
An end-to-end trading technology platform combining market data, strategy logic, automation, execution, analytics, research, machine-learning experimentation, and infrastructure. Built on behalf of Horizon Digital; all of the work is Ninef's although the repo commits sit under a separate GitHub account (statement S-9).
Highlights
- Develop an end-to-end trading platform (Python, FastAPI, PostgreSQL, Redis, Docker, Airflow, nginx) from market data and signals through execution, monitoring and analytics
- Built the automated trading path: TradingView webhook ingress, alert/execution logic and position monitoring, paper-trading options on Interactive Brokers
- Designed TradeVault, a Vue 3 + FastAPI trade journal (P&L, win rate, best/worst trades, calendars) that presents dense trading data without overwhelming the user
- Run ML and strategy research (MLflow, JupyterLab, Optuna, XGBoost, LSTM/PPO) across securities, timeframes, options strikes and Greeks on TSLA, NVDA, AAPL, QQQ, SPY, IWM and more
- Shipped internal dashboards and a React/TypeScript research UI behind a single nginx SSO gate with Firebase authentication
- Keep the architecture modular (Docker, MinIO, Superset BI) so services upgrade independently while backtesting, alerts, execution and analytics stay consistent