Taking pictures used to be a rare joy, now I have 30.000 pictures in my pocket. Unstructured, overwhelming chaos. We document our lives more then ever but seem to do very little with it. I’m building Memojo, a privacy-focused mobile app that runs entirely on device for both Android and iOS. The heart of the product is fast, offline vectorization that will support auto organisation, timeline/reel creation and much more. Everything must execute locally through quantized models, so experience with Core ML, TensorFlow Lite, Metal / GPU delegates or similar accelerators is essential. I already have baseline models tested and ready in our first prototype; what I need is a developer who can: • Package the models into a lean, cross-platform codebase • Expose simple APIs for text and image embedding calls • Handle on-device storage of the vectors and keep the pipeline fully offline • Build a clean UI that demonstrates both text and image workflows while I iterate on product design Acceptance criteria: a single code repository with Android and iOS projects, documented build steps, and a demo screen showing successful text and image vector outputs on typical mid-range devices. If you’ve shipped on-device ML apps before and enjoy doing more than Apple and Google combined, drop me a line!!