Compare the performance on asynchronous and synchronous architectures for Machine Learning deployment models and understanding where the performance delta comes from - View it on GitHub