Scaling ML Workloads

How distributed training and inference spread ML work across many machines to handle huge models and traffic.