Setting up an AI chat inference server involves selecting the right model architecture pair, provisioning hardware