We are looking for the development of a **face recognition component for a React Native mobile application** that works completely on-device. The scope of this project is limited strictly to face detection, embedding generation, and matching functionality. --- ## **Functional Requirements** The component must: * Accept **reference images/selfies** and: * Detect faces within each image * Generate and return facial embeddings for each detected face * Accept **media inputs (images and videos)** and: * Detect faces in images * For videos, process selected frames (e.g., initial or sampled frames) to detect faces * Generate embeddings for all detected faces * Provide **face matching functionality**: * Compare embeddings from media inputs against stored reference embeddings * Return match results with confidence scores * Be implemented as a **modular, reusable component/service** that can be integrated into an existing React Native app * Operate **fully within the React Native ecosystem**: * Can use native modules (iOS/Android) if required * Should expose a clean JavaScript/TypeScript interface * Handle real-world conditions: * Variations in lighting, pose, and image quality * Similar-looking individuals (e.g. cast members of a crew) * Family Members (e.g., siblings) --- ## **Performance Requirements** * Face detection, embedding generation, and matching should target: * **≤ 500 ms per operation** (depending on device capabilities) * Accuracy target: * **≥ 95% under typical usage conditions** --- ## **Technical Requirements** * Must use **open-source libraries** that are: * Free for commercial use * The solution should: * Be optimized for **mobile performance (iOS and Android)** * Minimize memory and battery usage * Work efficiently on mid-range devices * Preferred (but not mandatory): * Experience with libraries like TensorFlow Lite, ONNX, or similar mobile ML runtimes * Models that can be further extended to work seamlessly across both browser and mobile platforms developed using react. --- ## **Deliverables** * React Native-compatible **face recognition module** * Native bridge (if applicable) for iOS and Android * Clean API for: * Face detection * Embedding generation * Face matching * Documentation including: * Setup instructions * Example usage --- ## **Intellectual Property** * All deliverables will become the exclusive property of us. * The developer/vendor may not reuse or redistribute the solution ---