Over the last twenty-five years, Elsevier has transformed itself from a traditional print publisher into an information solutions provider. In the latest phase of this transformation, Elsevier is accelerating its efforts to combine the scientific and medical literature with a wide range of additional sources of data to create new ways of delivering value to customers in the research and healthcare markets we serve. The data science community at Elsevier, spread across our various businesses and embedded within our product development teams, is a key part of this effort. We will describe the organization and focus of our data science teams, the infrastructural and process-oriented challenges they face, and the principles they are applying to the creation of new data products. We will also provide a view into the range of data solutions being delivered to internal and external customers, and touch on a few examples of near-term and long-term data science projects, both of which exploit recent advances in automated knowledge base construction using neural approaches to natural language processing.