Suman Deb Roy

In my current role, I am a Data Scientist at Betaworks in NYC. My task involves building the popularity ranking architecture of Instapaper and Digg, data driven approaches to detect profitable seed investments and developing novel models of user patterns on the social web. I also work on building automated back end algorithms for growing startups to scale.

Being a combination of a scientist, developer and hacker, I love large-scale machine learning, data visualization, network analysis and predictive modeling that helps in interpreting the patterns and relationships mined from data to people in product development and marketing.

Before joining Betaworks, I completed my PhD in Computer Science at the University of Missouri-Columbia. My focus was on understand the vast distribution of data types on the Social Web, and how it led to observable phenomena. Having worked with enormous quantity and diversity of data found on the social web, I realized that the web is built of layers where information gets transferred via endogenous and exogenous factors. Two critical elements govern such information transfer - semantics and social networks. An algorithm that strives to understand, model and eventually automate such information transfer must be capable of learning these two layers at scale from many diverse domains on the Web. My dissertation, is as such, titled : 'On Cross-Domain Social Semantic Learning'