Inquiries About Internships

The short answer:My research group has a policy of not accepting short-term interns from outside of Carnegie Mellon.Sorry.Good luck in finding a suitable position.

Explanation:Every year, I get dozens of Email inquiries from foreign undergraduates about the possibility of spending a short-term summer internship with my research group at Carnegie Mellon.Lately, I have also been receiving inquiries about internships at other times in the year.

Iím not unique in this: all of my faculty colleagues get similar inquiries Ė in fact, we often get identical messages from the same students.Most of these inquiries come from India; a few come from China.Often the sender is from a good school and has good grades.Some even have a bit of research experience.

I like talented students, and donít want to discourage you, so for a few years I was sending individual replies to most of these inquiries.But I can no longer do that Ė it takes too much time.So from now on, I will simply reply with a pointer to this web page.

Most of these inquiries are generic form letters.There is no specificity: the same letter could be (and probably was) sent to hundreds of faculty members at dozens of universities.Often the letter starts by saying that the student is familiar with my research, is fascinated by it, and would love to work with me.Then, often, the note lists the studentís interests and experience Ė usually something like ďmachine learning, robotics, image processing, genetic algorithms, neural networks, theory, and formal logicĒ Ė things I do not currently work on, which you would know if you had looked at my web page.Finally there is usually a passionate statement that the student will work very hard and enthusiastically if accepted into my research group Ė but you didnít care enough to find out anything about my research, except maybe the name of the Scone project.Sometimes the English is OK, and sometimes there are indications that we will have real difficulty communicating.

Most of my colleagues just delete these generic messages without any reply.They consider them a kind of spam.I understand that itís a great opportunity to spend a few months with an elite research group at a top U.S. university.If you want to send out a lot of inquiries, OK.But itís unreasonable to expect a busy professor to spend more time replying to your message than you spent in sending it.

If your note indicates that youíve actually taken the trouble to learn something about our Scone project, and if you have reasonable preparation and some specific ideas or questions, you might get a personal reply from me.Iíve had some interesting technical discussions with students from many parts of the world.But when it comes to summer internships, the answer will be the same as the one above: sorry, but no.

There are several reasons for this:

      My group currently has no funding to pay for summer interns.If we did have, we would spend that money on Carnegie Mellon undergraduates, who are excellent, well prepared, and who might continue working with our group after the summer ends.

      Even if you require no funding, a 3-month internship would be a waste of time for both of us.At present, it takes a couple of months for even the best students to learn how to do useful work in the Scone system, so there is no chance that an outsider with no previous exposure to Scone wouldbe able to do anything useful during a 3-month visit. Experience has borne this out.
That may change in future years, once Scone is better documented and is readily available to outside users.But even in that case, I would want to see some clear evidence that you have learned about Scone on your own.

      Most people ask about visits in May-July.May is a very busy month here, with many end-of-year activities, and I am usually out of town for significant parts of June and July.I generally do not have a group of students working at CMU during the summer who could mentor visitors when I am away.

So, thank you for your interest, but I will not be able to offer you an internship with our research group.I apologize for not sending you an individual response.

-- Scott Fahlman

Addendum: I have recently returned to doing a little research on deep learning algorithms, just to see if my old work on Cascade Correlation can be updated to handle the larger projects of today. But this is preliminary work and I am not yet ready to build up a research group in this area or to take on any non-CMU interns.