CSIR India Jul 2024– Sep 2024 • Addressed critical need for rapid arsenic screening in rural wells; existing field devices were prohibitively expensive and inconsistent • Built end-to-end IoT solution: Arduino-based imaging rig with Android capture app, Python calibration pipeline, and ML classification system; optimized hardware bill of materials through iterative prototyping • Impact: Achieved ∼92% balanced accuracy on lab-validated samples; reduced prototype cost by ∼60% vs. commercial alternatives; tested 150+ water samples across villages; produced field documentation enabling operator handoff