Deploying Machine Learning Models to Ahead-of-Time Runtime on Edge Using MicroTVM