ML-Based Stock Decision Support System

Engineered a quantitative stock decision support system fusing fundamental and technical data to mitigate retail investment risks. Architected a decoupled Next.js and FastAPI infrastructure with automated ETL data pipelines for daily EOD market data ingestion. Implemented a Random Forest classifier optimized via SMOTE, successfully filtering out 126,000+ value traps. Delivered Explainable AI feature importance to ensure completely transparent, objective, and risk-adjusted trading decisions.