At Medical Informatics Lab, IIT Kharagpur, I built reproducible Python workflows for EEG and biomedical time-series analysis, from signal preprocessing and feature extraction to model benchmarking. I also translated neurological research hypotheses into measurable machine learning benchmarks with faculty and clinical experts.
My projects include MedVision AI, an explainable chest X-ray classification platform with Grad-CAM heatmaps and a FastAPI service, and a multimodal Parkinson’s disease framework that achieved an F1-score of 93.0% and ROC-AUC of 98.5% with regularized Logistic Regression. I’ve also developed medical classification methods and leakage-safe EEG models using subject-wise validation.

