Back to Projects
Professional Swing Trading System
A modular algorithmic trading system combining multi-factor signals, XGBoost prediction, SHAP explainability and LLM-powered analytics.
Overview
AI Swing Trader is an event-driven algorithmic trading system built around a modular architecture. It generates multi-factor trading signals, predicts price movement with XGBoost, and makes every decision auditable through SHAP explainability rather than leaving the model as a black box. Position sizing and stop levels are governed by ATR-based risk engines, while an LLM layer produces natural-language analytics on top of the quantitative output.
Key Achievements
Designed a modular, event-driven architecture separating signals, prediction, risk and analytics
Implemented XGBoost-based price movement prediction with multi-factor signal inputs
Added SHAP explainability so every trading decision can be traced back to its drivers
Built ATR-based risk engines for position sizing and stop management
Layered LLM-powered analytics on top of the quantitative pipeline
Tech Stack
PythonXGBoostSHAPEvent-Driven ArchitectureGenerative AILLMsQuantitative Analysis