Jetson Orin NX Offline Voice Assistant Developer - Fixed Price Milestones

I am building an offline AI voice assistant that runs completely on a Seeed reComputer J4012 with a Jetson Orin NX 16GB. No cloud, no internet, everything on device. The core stack is already chosen, researched and documented. All pre-built containers have been identified. I am not looking for research or discovery. I need a skilled developer who can take a well defined stack and bring it to life on real hardware. I am also open to developers who have already built something similar with compatible software. If you have a working solution that achieves the same result on Jetson hardware I genuinely want to hear about it. The finished assistant needs to wake on a voice command, understand quiet speech clearly at 1 to 3 metres in a room with background noise, respond in a warm defined AI persona, speak back naturally, remember the user across sessions and power cycles, and read content aloud on request. Hardware integration includes LED status light control with multiple behaviour states, clean shutdown and startup from a physical power button, and microphone agnostic audio that works with any microphone source without being hardcoded to a specific device. Fixed price milestones only. NDA required before any project documentation is shared. All intellectual property created during this project is owned by me as the client. To apply please include evidence of previous Jetson or embedded Linux work, your estimated timeline, and one sentence on how you would approach persistent memory across sessions and power cycles. Applications without all four of these will not be considered.