We are building a voice-activated AI wearable device — a compact, always-on personal assistant that listens for a wake word, processes speech on-device using a local AI model, and routes complex queries through the user's paired smartphone to cloud AI. Think hands-free information retrieval, GPS navigation, calls, messaging, reminders, camera capture, and object recognition and more — all from a wearable. The stack spans three layers: a custom Yocto Linux OS running on an NXP i.MX8M Plus, a Flutter mobile app (iOS + Android), and a Next.js cloud backend. The build system is fully working, the codebase exists across all three layers, and UI designs are finished. What we need is an experienced developer to bring the runtime to a working prototype — particularly the voice pipeline, which has latency and stability issues that need hands-on embedded Linux debugging. This role requires: Yocto / embedded Linux (non-negotiable) Python systems/service development Flutter (iOS + Android) BLE integration on both device and mobile sides This is not a greenfield project and not a finished one. It's a real mid-stage product with a clear technical roadmap. Serious candidates will receive a full technical brief including an independent code audit, complete architecture documentation, and a 30/60/90-day recovery plan before any commitment is made. If you have shipped embedded Linux products and mobile apps, we want to hear from you. NDA / IP are required.