Inference Optimization and Deployment: From Servers to Browser Runtimes | Advanced Deep Learning