An end-to-end PyTorch framework aligning chest X-rays and clinical reports using a ViT and Bio_ClinicalBERT. Features gradient-based visual grounding (XAI) and a multimodal RAG pipeline for context-aware radiological Q&A.

Fine-tuned a ResNet18 model using PyTorch for pneumonia classification from chest X-ray images, achieving 81.89% accuracy through model training and evaluation.
Applied attribution-based XAI methods, including Grad-CAM, Integrated Gradients, and Occlusion, to visualize model predictions and analyze image regions contributing to classification decisions.

My Portfolio --- This portfolio has been devloped using Nuxt v4 and TailwindCSS. By changing ts files in data folder, this project can be used as your portfolio.