– Remote Data Management System:∗ Collaborated with project stakeholders, including project managers, CEOs, and CTOs, to design the System. ∗ Designed and developed five interactive data collection form apps for field workers in ultra-remote areas, facilitatingreal-time advice and offline functionality.∗ Implemented AWS Lambda-based ETL data pipelines for efficient data processing and automated refinement∗ Engineered scalable database architecture for real-time data access and seamlessly integrated AWS QuickSight foradvanced data visualization.∗ Developed an ML-based high-risk prediction model to promptly alert ANMs on WhatsApp chatbot in real-time, thereby improving healthcare delivery in remote areas.∗ System is being used by The Heart and Stroke Foundation (South Africa), NICE Foundation (Hyderabad, India), andIDO (Kothagudem district, Telangana, India) and Successfully benefited about 5,000 patients.– Asynchronous Teleconsultancy System:∗ Led a team of three in designing and developing a SaaS-based Asynchronous Teleconsultancy System, leveragingLangchain. The system enables patients to report symptoms, submit images, inquire about FAQs, and access theplatform via a WhatsApp-based chatbot.∗ Crafted a symptom-based risk assessment model and image analysis tool employing Convolutional Neural Network(CNN) and transfer learningaccuracy of about 90 %. This tool efficiently identifies affected areas in patient images,facilitating ticket generation in the doctor interface with color-coded flags for criticality.∗ Integrated with hospital Electronic Health Records (EHR) systems for patient data retrieval∗ Currently undergoing trial at LV Prasad Eye Institute, Hyderabad, with over 1000 patients.

– Anemia Detection Model for eye conjunctiva:∗ Developed deep learning models: a u-net based segmentation model isolating eye conjunctiva and a neural networkbased classification model for Anemia detection, achieving 98.9% accuracy∗ Integrated into a chatbot for non-invasive Anemia screening among pregnant women at DMC Ludhiana as a trial withover 200 patient

I tried finding a corelation and later tried to search the reason for the that correlation