I did internship at IBM IRL Bangalore from 12-May-2011 to 05-Aug-2011 under Jayram T.S. who is the head of the algorithms department of IBM Almaden lab and is currently visting Bangalore.
I came to know about the internship from our revered senior Sreyash Kenkre. It seems that every year IBM sends pamphlets to Universities (including US universities) inviting students for internship. Even IIT Bombay got the invitation. I saw the notice on the KR Building Office notice board. But I agreed to go for internship only when Sreyash rebuked me. The process was simple. I sent my Resume to Jayram & Vinayak Pandit of IRL. They took an telephonic interview (when IND-PAK semifinal was going on). Mainly they talked about the project and then I talked about my work. After around 2 weeks they sent me an offer.
The work was totally new for me and was not related to my Thesis. They wanted to explore about the possible way of solving a problem on linear algebra using different techniques. For that I needed to do some coding. Mainly I needed to solve a large number of LPs using Matlab/C++. But the main part was to explain the observations using theory. For that we needed to go through some papers on unique solutions of LPs / Counting the faces of polytopes etc. Finally we realized that the problem is related to the field of compressed sensing.
It was a very nice experience working with Jayram. He is a good teacher. Whenever there were some terms/topics unknown to me in the papers we were reading, he would explain me the basics of them himself instead of asking me to read about them offline. Overall the feeling was that we were learning together intercepted by he teaching me the basics.
|In front of security desk of IRL. More photos here.|
I think PhD students should go for internships (may be in the summers of 2nd and 3rd year) which will enable them to expand their horizon. Since I had prior work experience, I thought I did not need to go for internship. But now I would like to say that working with research lab is a different experience. In the worst case, it will help us decide about what to do after the PhD. (Most of my co-interns from US had already done internship in the summers of previous years.)
Another thing I realized that we (theorychat walas) should discuss more among ourselves about new papers/topics even not directly related to our thesis. For example we can make theory chat to be something like this: Dartmouth Theory Reading Group