At CannazoIndia, I analyzed inventory and operational data to identify inefficiencies and support process improvements that streamlined day-to-day operations. I also managed two subordinates and worked with two business partners.
On MyCart, I built a multimodal agentic RAG shopping assistant with an eight-node workflow, conditional routing, and grounded response generation. I added image-based product search and integrated the chatbot into a Django e-commerce platform.
For my Loan Default Prediction project, I prepared applicant data, engineered predictive features, and addressed class imbalance with SMOTE and balanced sampling. My Random Forest model achieved 89.85% accuracy.
