Bellevue, Washington, United States
I work in Bing's Indexing and knowledge team which is responsible for processing and extracting all sorts of interesting things from each document in Bing's index. Most recently I conceptualized and developed the end to end, data driven pipeline for fact extraction from semi-structured (like Wikipedia infobox) and unstructured data (like natural language sentences). It is one of the fundamental sources will power the Bing's entity snapshot experience[1]. Specialties: Machine Learning, Statistical Natural Language Processing, Information Extraction [1] http://www.bing.com/blogs/site_blogs/b/search/archive/2013/03/21/satorii.aspx
Developed a generic Document classification systems with the ability to learn from a minimal amount of supervised training. Concrete implementation for handling Documents like emails, user activity messages, chat messages and tasks was done. Incorporated an existing task identification library for flagging Documents containing tasks for highlighting and routing.
Worked on the development of Stock Exchange trading terminal for nearly 3 years. Worked with a multitude of programming languages like Java, C, C# and Visual Basic. Implemented various standard and proprietary network protocols like Financial Information Exchange etc. Conceptualized and implemented a UDP message filtering layer for a very high volume UDP gateway. This resulted in reduction of idle CPU utilization from over 95% to under 5%. Consistently got a performance rating of 4 and higher, on a 5 point scale.