On Retail-OS, I architected a 20-agent retail platform that processed 86,000+ records across 15 data sources. I built an XGBoost demand forecasting model and anomaly detection systems as part of the platform.
I also developed a stock market analysis and ML trading system using NSE India data. Its Q-Learning agent improved returns from −18.8% to +3.4%, and its Hierarchical Risk Parity portfolio achieved a 124.3% return while reducing maximum drawdown compared with the equal-weight benchmark.
For my Smart Traffic & Gate Access Management System, I built vehicle access control with YOLOv8, OpenCV, OCR, and CNN-based vehicle classification, integrated with an SQL database. I also deployed gate control using an ESP32 and servo motor.
As a Product Development Intern at VDC (Venture Development Center), I designed 3D mechanical enclosures in Fusion 360 and built hardware-software prototypes for smart access management and pet feeding. At GITAM Bengaluru, I conducted ML training sessions on feature engineering, model validation, and model selection.
