To solve real-world problems, it's sometimes necessary to run computationally heavy models. Properly leveraging parallel processing and other optimizations can save precious time and resources for training and inference. In this session, we’ll introduce NP on MXNet, a drop-in replacement library of official numpy for accelerating your workloads and opening your models to even more possibilities. The session will introduce the key features of NP on MXNet and give a demo how to build a machine learning model with this new package.