A routine visit to the dentist should not end in a life-or-death emergency. For me, it did. A critical pharmacy error, swapping my prescribed antibiotic for a potent diabetes medication, sent my blood sugar plummeting to a near-fatal 2.7 mg/dL. I was rushed to the ER, barely conscious, the victim of a preventable mistake that statistics show affects millions annually. Lying in that hospital bed, I wasn't just thinking about my recovery; I was thinking about how technology could prevent this from ever happening to anyone else. In this session, I will share my personal journey from a visit at the dentist over hospitalisation to building a functional prototype that leverages the power of multimodal Large Language Models (LLMs) like Google Gemini and MedGemma. I will walk you through the initial concept of MedBuddy, from the frustration and fear that sparked the idea to the design and development process. We'll explore how MedBuddy allows users to simply take a photo of their medication and prescription, and the AI will analyze the visual data and text to verify if the medication is correct. More importantly, I'll share my vision for a future where technology empowers patients to take a more active role in their healthcare, preventing the kind of life-threatening mistakes that I experienced. This is a story of how a personal disaster was transformed into a passion for creating technology that can save lives.