If you arrived at this page, chances are you are either writing your thesis with me, or at least interested in doing so in the future.

Perspective students#

In general, I am happy to entertain any topic that we together come up with, provided it falls in at least one of these categories:

1. Natural Language Processing / Computer Vision / Machine Learning and their applications

These are the things I am academically interested in. Particularly the "applications" part.

If you manage to find a topic that fits in here, chances are high I would be happy to supervise your thesis.

2. Fun / Strange / Crazy

It is a deeply held belief of mine that the current "industrialized" academic environment has lost something of its founding narrative: it was supposed to be a place to try and test new ideas in a playful manner, no matter how strange (or even crazy) they may sound.

If this sound like something you would agree with and your thesis would fit, do reach out and we’ll certainly find a common ground.

3. Useful Open Source Software

Open Source Software is what much of modernity has been built upon, no matter where and how you look. If there is a chance your thesis will produce some Open Source Software that has a shot at being useful, I am sure we can figure something out together.

In any case, please do reach out to me as soon as possible -- you can find the contact details towards the bottom of the page.

For those already "in"#

1. Setup the writing platform#

Make sure as soon as possible so that we set up a platform together where the writing will happen. I am very happy with a shared Git repository but a shared Overleaf project works too.

2. Read "Write the Paper First"#

Read the "Write the Paper First" article by Jason Eisner. It’s from 2010, and so some of it might seem a bit dated, but overall it’s one of the best pieces about academic writing I came across that also serves as a nice organizational framework.

We will be making use of it too.

Unless the goal of your thesis is to build specific software to specs, there is a significant chance you’ll spend the first 20 to 30 percent of your time learning about and reading upon "Related Work", the body of (often academic) literature that touches upon the concepts relevant to the thesis.

When you do, the best piece of advice you can get is to take a lot of notes --preferably in the document (or one of the documents) we’ve already set up together. Any time an article or a book you are reading answers a question you might have had, put both the question as well as the answer down in the document.

Here are two guides that I found helpful:

The reason for all this is quite simple. The "Related Work" section will constitute a significant chunk of your thesis text, which is often very tough to write in retrospect. Having it all in one place, no matter how disorganized, will save your a ton of time (and nerves!) in the long run.