My name is Mohit and I currently work at Pandora as Scientist. I mostly work on solving recommendation and personalization problems.
I enjoy solving large scale machine learning problems. I have developed a particular liking towards using (and learning) non-parameteric bayesian techniques. I find the idea of writing a "custom" graphical model to solve a particular problem very attractive. I have also used randomized techniques to solve large scale linear algebra problems. Recent advances in Rand-NLA (like CX/ CUR factorization) have demonstrated how randomization can be used to solve TB size problem. In the end, what motivates is deploy an algorithm in the production to see it make an impact across millions of end users.
On personal end, I enjoy running, watching movies and traveling. I am also involved with PugPros Rescue in an attempt to make some difference in lives of our furry friends.