@Irfan-Usman I don't work for ModalAI but a few notes that might help you.
https://gitlab.com/voxl-public/voxl-sdk/services/voxl-tflite-server is a good starting point -- it runs Tensorflow Lite on the VOXL/VOXL2 directly and has some built in models such as MobileNet and YoloV5. https://gitlab.com/voxl-public/voxl-sdk/services/voxl-tflite-server/-/tree/master/misc_files/usr/bin/dnn
I am doing similar with ModalAI's platform -- you can reach me through the email in my profile.