Breast cancer is a leading cause of cancer mortality among women globally, with disproportionately high late-stage detection rates in low-resource rural settings. This paper addresses the critical need for accessible, low-cost screening infrastructure in rural areas of Bihar, India, where geographical, economic, and socio-cultural barriers (including significant stigma) impede timely diagnosis. We propose a robust, two-tiered Machine Learning (ML) framework designed for Edge Computing deployment. The framework utilizes two models: 1) A Deep Convolutional Neural Network (CNN), fine-tuned using Transfer Learning on the public CBIS-DDSM dataset, for the binary classification of suspicious lesions identified in imaging/scanning data. 2) A Random Forest Classifier trained on structured clinical and demographic data to provide personalized, localized breast cancer risk prediction.