Thursday, August 09, 2012

Advice for an Undergraduate

I have been asked by an undergraduate student to help him out on finding a PhD studentship in Compressive Sensing. I am not sure I am the right person to answer that question but here is what I know:

  • You need a LinkedIn page preferably one that has an English version
  • You need a webpage where you can tell the world about even the most insignificant technical projects you have ever undertaken. Both your webpage and the LinkedIn page need to be cross-linked. You might even consider starting an account on Github. 
  • If you are able to do so time-wise, you might even consider blogging on technical issues. An impressive example of somebody who has done well in this regards is Danny Bickson. Danny was making his code available when he was a student, he then started a blog when GraphLab needed to be explained to newcomers and he just organized a workshop with more than 300 people from the Silicon Valley and other Fortune 500 companies who now know who he is and probably want to work with him.
  • If English is not your first language and you have no way of getting corrections from a friend, I suggest you come back to your past blog entries or web pages often and correct them for style. 

All these things are the seeds of a conversation. and only that. To get that conversation started, it's a little bit like flirting:

  • You need to interact directly with potential interested PhD supervisors. Face-to-face is better than e-mail. A friend of mine got into MIT some years back by coming to the department he was interested in a year earlier on a tourist visa. He visited his future PhD advisor who in turn invited him on an unpaid internship a year later in the Summer.. When you are in a room and tell somebody you crossed an ocean for him or her, trust me the conversation gets going. 
  • For most it may be too hard so the next best thing is to start conversations. How do you do that ? well kindly interacting on the LinkedIn Compressive Sensing (1683 members) or the Advanced Matrix Factorization group (388 members) would be a good start. Be gentle, people are busy, you cannot start a conversation with " I need to find all the references on Compressive sensing and face detection" or " I want to do a PhD in Compressive Sensing in Wheat genomics". You don't start a conversation like this when you are flirting, why should this be case here ?
  • Typos: If you want to send the signal to a specific person that you really care about their work, read through the preprints they have listed on their website and find the typos. Send them an e-mail about it. It's a great conversation starter.
  • If you have already done something that is even remotely relevant to compressive sensing like using an Arduino and a Kinect to produce a calibration mechanism for a random imaging system or implement a simple algorithm in Octave or run different algorithms on a set of benchmarks or produce a new sets of CS examples on scikit-learn, please do let me know.  As I have done in the past, I'll be the first to feature it on Nuit Blanche. To have an audience, trust me, the hardest part are the first million three hundred and thirty three thousand page views. We are already past that here.

Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

No comments: