Thread by @vj_chidambaram

A quick thread on working with students remotely. As the saying goes, "talent is uniformly distributed, but opportunity is not".

I've had success working with indian undergrads remotely, and found some of my awesome students via this approach: @RohanKadekodi and @Ponnapalli95

@RohanKadekodi @Ponnapalli95 To provide some background for this: indian grad students not in the IITs or NITs have low exposure to research, and very few opportunities to do research in their undergrad. This is despite many students being super smart, motivated, and hard working.

@RohanKadekodi @Ponnapalli95 This is a pool of students that is largely untapped by American grad schools. However, the pool has large variance in the quality of students, and it takes significant effort to find the good students.

I've been working with students remotely over the past three years to identify and recruit these students. This is how I found @RohanKadekodi and @Ponnapalli95 (or rather, they found me): both worked with me remotely for about six months before joining my group.

I'm also working remotely with @hubatrix through the ReportBee Fellowship, and we had a HotStorage workshop paper out earlier this year.

cs.utexas.edu/~vijay/papers/…

I worked with Nikhil Sambus remotely and he went off to the University of Toronto to work with Prof. Ashvin Goel :)

These days, when a promising student emails me, I offer to work remotely with them to strengthen their application. I make it clear I'm not looking for students myself, but I would be happy to work with them and write them a letter if things go well.

About 90% of these interactions fizzle out after the first few emails, so its not a lot of effort on my part. But the students who remain tend to be motivated, organized, and hard working since they are doing all this on the side on top of regular school work.

I highly recommend my fellow assistant profs to try this. This is an awesome new pool of students. Its takes effort to find the good students, but its usually worth it since there is a very good chance the student will come and work with you during grad school.

Here's my low cost method for doing this. Open-source your research prototype on Github. Have issues you tag as "Help Wanted". Direct students who want to contribute to these issues. See who actually codes and ask questions.