At TUM, topics for available thesis projects are advertised on faculty web pages. I expect the situation to be similar in other universities.
I refer to the collection of open thesis topics as the thesis topic market. The topics on the web pages are its visible part. However, there is also a hidden topic market that contains topics that are available but never get advertised.
The Hidden Topic Market
Of the 25+ thesis projects I supervised to date, only 2 students found me through the visible topic market.
I simply don’t advertise most of my topics. Why? Because it does not pay off. Advertising a topic requires time for producing a polished description. However, I typically don’t get much response—or only from students who did not even take the time to proof-read their application emails; I have received emails with 3 sentences and 8 errors. If they don’t even care to proof-read their application, how much care will they invest in their thesis?
It works much better for me to only advertise topics to individual students. Either because I met them during a teaching assignment, or because they approached me directly. As far as I know, many other supervisors do it alike.
Cracking the Hidden Thesis Topic Market
The key to cracking the hidden topic market is to show initiative. Simply approach potential supervisors and ask them for topics they might have available.
Why does this work? Firstly, initiative from a student is a great indicator for good performance in a thesis project. Since every supervisor prefers to supervise students that produce great theses, showing initiative makes you more likely to be accepted. Secondly, as a PhD student, few people care about one’s work. If anybody approaches you and shows a little interest, that person is likely to get your sympathy.
First, find candidates for supervisors:
- Ask your friends to suggest good supervisors
- Go through the courses you took: do you remember any teacher that you liked and worked in an interesting area?
- Browse completed theses. Any area or topic that catches your eye? Maybe there are similar topics available in the same group, possibly building on this work.
- Finally: browse existing thesis topics in the visible market. Again: even if a topic is not a perfect fit, their supervisor often has similar topics.
Second, get rough understanding of their background. In which area of computer science are they interested? To which topics do they publish? Which position do they hold (e.g. PhD student, post-doc or professor?). Don’t exaggerate: investing 20 minutes in learning the background of a supervisor candidate probably puts you way ahead of your fellow students.
Third, contact the most interesting candidates. I suggest writing a short, well-written email. Make sure the email does not read like a bulk mailing that you sent out to other supervisors, too. You could e.g. include these points:
- How you came across that person (e.g. personal reference or thesis topic or thesis supervised before?)
- What topic or area you are interested in (ideally referring to his work)
- What makes you an interesting pick for him (maybe you have taken a course in this area? Or have experience in the programming language required for the project?)
- Aks for the possibility to meet in person (e.g. for coffee)
Finally, the worst that can happen is that they turn you down. You are then in the same position as before. There is thus little to loose in contacting supervisor candidates.