I am a Natural Language Processing (NLP) researcher. NLP is a field in Artificial Intelligence that aims to understand and generate text. The most popular application areas in NLP include chatbots, automatic language translation, recommendation systems, etc

Legal Natural Language Processing

At Thomson Reuters, I work on legal documents, applying machine learning and NLP to extract useful information for legal experts. The most challenging part of legal NLP is that the legal text is not like traditional text like news, social media, or Wikipedia. So, my teammates and I build customized legal AI systems with the help of Subject Matter Experts.

Tamil Natural Language Processing

I spend most of my Sunday afternoons working on Tamil NLP projects. Be it building Tamil NLP datasets or developing machine learning models for Tamil text, I am always excited. Apart from enjoying Tamil NLP research personally, I also see it as my responsibility to give back to where I come from. I collaborate with fellow NLP researchers across the world who are interested in Tamil. We have tonnes of ideas and always are looking for motivated researchers to collaborate. Please get in touch if you are interested.

Extracting Possessions

My entire life was about extracting possessions during my PhD. Possession is a semantic relation between two entities where one entity possesses another entity. To simply define, I looked for who owned (or rents or is in control) what. After identifying what a person possesses, I identified when the possession relation is true (before, during, after), and how long a possession relation holds (ice cream – 1 hour? car-3 to 5 years?). I also identified if multiple people can possess one entity (a refrigerator at home). I also generated possession timelines for art works (Monalisa, Guernica, etc.) Read my doctoral dissertation on possession extraction here.