Increasingly, data-science applications require the creation, manipulation, and analysis of large graphs ranging from social networks to language models. While existing graph systems (e.g., GraphBuilder, Titan, and Giraph) address specific stages of a typical graph-analytics pipeline (e.g.
