AIT Application Guidelines

If you want to become a PhD student in the AIT group please take a couple of minutes to read the below. Here are a few facts and ideas that will save time for both of us.


I am always looking for really good graduate students to do interesting new research. The long-term partnership between student and supervisor is one of the most rewarding aspects of doing research in academy as opposed to industry (I've tried both).

Keep in mind though that this only works out if enough of the below constraints are fulfilled. When I say really good - then I really mean it. ETH is one of the best Universities in Europe and the world - many ETH students (and PhDs in particular) have always been at the top of their class at whatever institution they come from. If your previous track record isn't outstanding and - more importantly - you don't think you will become the best of the best at what you plan to do during your studies, then don't apply! You will feel out of place, you will suffer from peer pressure and you will suffer from the weight of expectations (to perform at the highest levels).

On the flip side - if you always have been at the top of the pack, if you always strive to do top quality work and learn challenging new things every day then please proceed.

This is not meant to say you need to be arrogant; you just need to be really good, and you need to be prepared to interact and cooperate with other really good people. People that are often lot smarter than you. Even if this rarely happened in your previous school, it will definitely happen here. Happens to me, too, of course. And it's a nice, useful and humbling feature of being at ETH or any other top University.

What kind of person do we look for?

Someone good! Show me good grades, awards, references, and so forth. Someone who likes to build stuff - and doesn't shy away from complexity! Show me something you've build and that you're proud of; something that works and that is novel and interesting; if real people are actually using it, even better. I like hackers, builders, makers and creators! Someone with some idea of what they want to do here (show me a good research proposal). Someone interested in something I will find exciting and fun to work on (read our webpage for hints, but then come up with something new on your own). Last but not least, someone with an interest and appreciation for aesthetics - this might be surprising for a PhD candidate in CS but HCI is very broad and interdisciplinary. You need to at least be able to distinguish good from bad design (you don't need to be a designer or artist though).

In general if you have these qualities you will do well in our group:

Are quick, smart, creative, sociable, outgoing, and funny
Enjoy working with others in a group
Past performance
Finished their Diplom/MSc at a top university, with outstanding results (Within the top 5% of the GPA at your school).
General skills
Are very quick at picking up new information
Fulfil the "first-derivative of learning" rule: In the end, being quick at learning new knowledge is much more important than their current level of knowledge (for example in HCI)
Have a self-driven interest in uncovering and solving unknown problems
Are able to work hard and creatively without constant supervision
Fulfil the "fire-and-forget" rule: I love working with people who go off with a task without coming right back with questions that I think they could have easily answered themselves (Google is your friend).
Domain knowledge
Have good programming skills (did I mention that I like systems builders)
Ideally have a CS and HCI background, ideally during BSc+MSc / Diplom studies
Ideally know the basic introductory HCI literature.
Have a sense for good and bad user interfaces and how to create them (the good ones)
Research skills
Have already published a peer-reviewed conference or journal paper as primary author, ideally at UIST, Siggraph, CHI, or a related top-tier conference in Computer Vision, Machine Learning, Robotics...
Teaching skills
Have worked as teaching Hiwi or in other teaching activities
Enjoy teaching, working with, and advising students
Language skills
Speak, understand, and write English fluently
Speak and understand German, or are willing to learn

How do you stand out?

We get lots of applications - all the time. In fact so many that I could spend entire days looking through all of them, let alone answering to them. So how do you make sure yours gets noticed? First, make sure your profile matches the above description and the research areas that we operate in (again browse through our webpage). Second, the sure fire way to catch my attention is to write a good research statement! If you show me that you already understand what is worthwhile research and that you have good ideas in a relevant space I will be all ears. Furthermore, a good research plan shows that you understand the area well enough to point at interesting challenges. Make sure you can communicate why your idea is interesting, how it is unsolved, and what your insight is that makes it solvable in three to four years.

Now the key to writing a good proposal is topic selection. The secret is to select a topic that few have yet identified as a research problem. If you pick something that everyone and their mum have studied to death, then it is so much harder to make a contribution. On the flip side if you pick a topic that is significantly novel there should be plentiful opportunities to make a contribution. If you want to discuss your idea with me before the real submission then please be clear, to the point and concise. If you can't convey the core of your idea in fewer than 300 words, probably the idea isn't really clear to you either.

What are (y)our goals?

This one is easy. Our group strives for quality and excellence in all aspects of research. If you want to go into academia or industrial research lab after your PhD then your goals should be aligned with ours. That is working on interesting and challenging problems, coming up with novel solutions and pushing your technologies to a level where real people can use them - and finally publishing your work in top tier venues. If this sounds like you then we will get along just fine.

So in summary - if you think you have what it takes to get a PhD at ETH and in our group. Then go ahead and send us your application (and research statement).

Master's program

If you’re interested in doing your MSc first you will need to apply centrally with ETH. More information can be found on the department’s website. Once you’re admitted and want to get a head-start in doing research come talk to us for projects and thesis work.


Not really. Generally speaking 3 month internships are too short to get anything significant done – especially if you are new to the area. Unless you come with personal recommendation from your advisor (that I know) and unless you have already done some interesting work in the area of HCI I will not even answer emails about Internships.