A personalized AI-powered platform designed to help my sister prepare for her medical specialty interviews. The system simulates realistic interview scenarios with automatic voice-based Q&A, provides detailed AI feedback and sentiment analysis and ensures exam integrity by monitoring for tab switching and suspicious behavior using the webcam and screen recording. This tool delivers a secure, supportive, and effective interview practice experience tailored for medical professionals.

Note:

• Webcam monitoring didn’t work with screen recording

• Echo in questions is from mic picking up speaker sound, won’t happen in real use 

An enterprise-grade AI pipeline for Systems Limited that automates health insurance reimbursement by extracting handwritten prescription data using Azure OCR, enriching it with Azure OpenAI and validating each entry against internal health policies through semantic reasoning and business logic rules. The system handles unclear handwriting, flags errors, and exposes results through a secure REST API and React front end, enabling the Medical claim team to process reimbursements more efficiently, accurately and in compliance with policy standards.  Show More

A conversational shopping assistant built in React with Azure OpenAI that lets users browse products, manage their cart, and complete purchases seamlessly within a chat interface styled with Tailwind and powered by Vite. Modular React components handle the chat flow and product cards, while backend JSON state and OpenAI intent parsing enable a smooth, AI-enhanced retail experience from browsing to checkout. 

A text based ML pipeline using scikit‑learn to classify news articles as real or fake featuring TF‑IDF vectorization, model comparisons including Naive Bayes, SVM, Random Forest and others, and comprehensive evaluation via accuracy and confusion matrices. Deployed with a Streamlit app and Flask /predict API, the system handles short or ambiguous headlines and offers real‑time interactive predictions