๐ I work as a Research Engineer at Google DeepMind India. I am part of the Languages team led by Partha Talukdar.
I received my master’s degree in CS from IIT Bombay. My thesis (advised by Prof. Sunita Sarawagi) was largely about methods to improve quality of text generated from language models. These methods were primarily applied to adapt language models to new domains or reduce hallucinations on unseen domains.
I also spent some time at Microsoft Research India with Sunayana Sitaram working on multilinguality, evaluation methods and cultural aspects of LLMs.
My broad research interests are domain adaptation, machine learning for low data regimes and machine learning for (computer) games.
One of my favorite video on AI is this lecture by Prof. Patrick Winston on the strong story hypothesis. The lecture also presents an interesting take on human intelligence being a “discovery rather than evolution” which I (currently) agree with.