TL;DR: I am holding office hours to help (undergraduates, masters) students get started with research.

(Tentative) time: 9:30 AM EST or 3 PM EST (every Friday). Location: Virtual (fill this short form, and I’ll send you the link). Long version: Six years ago, I began to orient my career towards research. This was unpopular in my peer circles because my undergraduate program offered little to no research opportunities. My situation was aggravated by the fact that no one in my extended family considered research as a viable career, and only a handful of my family members knew about a Ph.D. To them, the idea of “studying” for another 5-6 years when one could find a job immediately simply didn’t make sense. Fortunately, I found seniors (mostly strangers at first) who answered my questions from time to time, encouraged me, and suggested people who I could work with. Thanks to many lucky breaks early on, right now I am in a privileged position. Every day, I marvel at the opportunity of being able to do research alongside brilliant faculty and students at CMU. Occasionally, I receive emails from students who are at a stage similar to where I was six years ago. Their questions remind me how lucky I have been, and how difficult it could be to get started with research.

In a small effort to make research more accessible, I am considering dedicating some time every week to conduct office hours virtually. The office hours are primarily meant to discuss i) general concerns pertaining to research, or ii) specific research directions related to NLP/ML (my expertise lies around interpretability of neural models). The general questions could span whether you should do a PhD, how to build a strong profile for graduate school applications, how to initiate collaborations, etc. There is a lot of good general purpose advice on the web, but very little of that is personalized. The office hours aim to offer guidance on your specific case. If people are interested in discussing similar topics, we can organize small group sessions, else we’ll resort to 1:1 interactions.