Today, we are facing another industrial revolution through machine learning technology, which would radically change our daily life in near future. We might see everyday news articles that say artificial intelligence can do something like human e.g. playing Go, classifying images, recognize speech, drawing Gogh, composing music, etc. Some of them already work well enough to be commercialized, but many of them are just prototypes and still needed to be improved for real world usages. In the era of deep learning, developing machine learning algorithm becomes easier with help of well-developed tools when comparing to traditional hand-crafted algorithm development. However, it becomes more difficult to develop commercial grade product due to large amount of data and poor interpretability and tunability of deep neural networks. In this talk, I will introduce necessary technology components to ease the difficulties and our vision to help everyone to solve real world problems.