Inference Server

This is what runs on the GPUs and handles JSON input, runs whatever logic you want in Python, then returns JSON output.