Hidden Wins in Data Engineering

Data engineering forms the platform on which all of data science magic happens. Having a solid data infrastructure tremendously boosts the performance of decision science, business analytics and machine learning and results in faster product iterations. In this talk, we will share some of the unconventional learnings we gained while setting up the foundation of data science from our experience at Coursera. Data engineering is a superset of business intelligence, data infrastructure and data warehousing, and also forms the base of all data science and machine learning work. The projects in our team have ranged from ETL, data standardization, data warehousing to building user facing services for recommendations and A/B testing. We have repeatedly observed how certain patterns have made it easy for us to maintain a data driven culture at Coursera. In this talk, we'll cover various topics such as benefits of unidirectional data flow, building internal tools for data accessibility and democratization, and the need for data standardization to improve data usability. The audience can take away key learnings to implement in their organization to help boost productivity of their data science teams as well as steer the company towards being more data driven.