Isha Gupta

I've built machine learning and computer vision systems through internships at Samsung PRISM, IIT Roorkee's Mehta Family School for Data Science and Artificial Intelligence, and TIET's Experiential Learning Centre.

At Samsung PRISM, I developed machine learning models for enhancement assessment in RAW DNG image buffers and built an end-to-end Python ETL pipeline for DNG decoding, RAW extraction, image enhancement, and quantitative quality assessment.

At IIT Roorkee, I researched differential privacy and membership inference attacks, implementing Python simulations to evaluate the robustness of privacy-preserving machine learning techniques. My work on privacy-preserving neuroimaging analysis was accepted at a conference organized by Sejong University and selected for Springer LNNS publication.

I'm also building applied AI projects including PrivaScanFL, a federated brain MRI classification pipeline, and AgniPathik, an AI-powered smart fire evacuation system. Earlier, I programmed Arduino-based motion control, navigation, and obstacle detection for an autonomous cleaning robot.