Aryan Kapoor

Developing an AI-powered healthcare chatbot and clinical decision-support assistant that analyzes patient records, diagnostic reports, and longitudinal medical history to generate context-aware insights for physicians. Architected a scalable OCR and document intelligence pipeline to extract and structure clinical data from medical reports, prescriptions, and diagnostic documents. Designed and implemented a hybrid Retrieval-Augmented Generation (RAG) architecture combining Graph RAG for structured patient data with vector-based RAG.

Developed Cyber Shield Pro, an AI cybersecurity assistant leveraging agentic RAG, multi-step reasoning, and tool orchestration. Built a scalable NLP media monitoring pipeline with 11 models (OCR, translation, summarization, sentiment, classification) using Mistral-7B, IndicTrans2, mBART-50, optimized via vLLM. Developed a Conditional GAN (CGAN) with Wasserstein loss (WGAN) for spatio-temporal trajectory data imputation.

Architected a recommendation system using collaborative filtering and content-based methods. Built and evaluated a sentiment analysis model for customer feedback classification, achieving 89.45% accuracy. Developed an AI-powered customer support chatbot using NLP techniques.

Optimized a CNN model for medical diagnostics through hyperparameter tuning and transfer learning, achieving 93% accuracy. Conducted thorough data analysis and cleaning on 10,000+ medical images.