Deep learning model developed in collaboration with Stanford faculty to enable multi-level pneumonia classification in under-resourced hospitals. Optimized for deployment constraints including limited computational resources, minimal infrastructure, and low-bandwidth environments.