Uber机器学习平台Michelangelo技术负责人及联合创始人。负责training和serving系统。加入uber之前，曾在谷歌和通用电气参与地图，搜索引擎和医学图像的数据处理工作。Eric Chen 是滑铁卢大学博士，专注于大数据几何问题。
In Uber, we build one machine learning system to serve all machine learning needs across the company. On training side, that requires a system to train models with different workflows, different scales, different data properties, and different data localities. On serving side, that requires a system to serve models in online and batch applications with scale and consistency. In this talk, we presents how we build such a system. We will also demonstrate an NLP use case about how we build and serve models using combined Deeplearning and non Deeplearning techniques.