I’m building “CalorieSnap,” a cross-platform mobile application that lets people snap or upload a photo of any meal, instantly see an estimated calorie count, view a macronutrient breakdown, and save the entry to a personal food log. The goal is a fast, intuitive experience that works equally well on iOS and Android. Here’s what I need you to deliver: • A polished mobile app for both iOS and Android with seamless camera capture and gallery import. • An on-device or hybrid AI pipeline (TensorFlow Lite, Core ML, or a comparable framework) that recognises common foods, estimates portion size, and returns calories plus basic nutritional insights. • A persistent food log where each analysed photo is stored, editable, and totals are aggregated by day. • Optional but welcome: a lightweight backend (Firebase, Supabase, Django, or Node) to sync data across devices and enable simple account management. • Full source code, model files, build instructions, and a short screen-recorded walkthrough. Acceptance criteria • Calorie estimation accuracy within ±15 % for at least 1,000 popular dishes. • Camera and gallery workflows complete in under five seconds on a mid-range device. • Logged meals can be created, updated, and deleted, with daily calorie totals visible at a glance. Share any relevant computer-vision or nutrition-tracking projects you’ve delivered, the stack you prefer, and how long you’d need to get to an MVP. Let’s create something that makes healthy eating effortless.