How to Write a PhD Motivation Letter

(For Me)

I’ve been faculty for just over two years now. After reviewing a few hundred motivation letters for PhD applications, I’ve noticed some patterns, and often wished I had a chance to tell applicants the following things. I don’t claim that this advice is good for applying with other people, but it’s what I want to see, and my close peers tend to agree.

In a PhD application, the documents that interest me most are the motivation letter and the CV, but arguably the motivation letter is by far the most informative. It is the only document that can give me information about how our next four years working together can look like, and this is what I look for.

In short, this can be summarized as three main points: 1. Why are you interested in this topic? 2. Why are you interested in working in this lab? 3. What would you like to do?

Don’t get me wrong: you don’t have to have fantastic answers for all three. But anything you can come up with for each of these points will give me good insight into how we would work together.

Caveat: All PhD positions in my department require a thesis-based (research) Masters’ degree or equivalent. The PhD lasts 4 years and is more project-driven than is usual for instance in the US. Hence our recruiting process and expectations may look different than what you are used to.

  1. Why this topic?
    Acknowledge the topic and show your interest in it.
    A PhD with me is a research position in machine learning for natural language processing. This may sound weird, but just acknowledge that you understand what this means in a sentence or two. If the job is on a concrete topic (e.g., machine translation), just briefly say what the topic means to you, and why you care about it. If the job is more open-ended, pick one of the directions and do the same. Be concrete here.
    👎 “Machine translation is a very important technology that impacts people’s lives. I have been interested in artificial intelligence ever since high school and in particular I am highly motivated to work on machine translation.”: insufficient: without additional evidence, i can’t even be sure you know what machine translation means.
    👍 “Machine translation can empower migrants with access to important resources, and I used such services often at hospitals while living in Lisbon, for instance. However, I’m lucky to speak English, because translation into lower-resource languages like my native language remains substantially worse even for the large deep networks in use today”: much better, shows you know what it is, personal connection to the topic, and some awareness of the “state of the art” in the field.
  2. Why in this lab / with me?
    Justify why you’re interested in this position in particular.
    A PhD is a fairly long commitment, and you should not apply to work anywhere and with anyone without looking into them and the lab a bit. If you did this and you’re still interested, share with me why! If you have any possible doubts, share them in the letter too. This does not have to be about research only, but also about values, vision, personal fit, etc.
    👎 “The University of Amsterdam is a leading institution that publishes many papers, so I believe it would be the best place for me.” Not bad (better than getting the university name wrong), but if it’s basically interchangeable between all your applications, it doesn’t give me any insight into what you want and why.
    👍“The Language Technology Lab has been consistently doing research in machine translation that I have enjoyed reading, for instance the work of Name (2021) has motivated me to work in this field.” — specific, good.

  3. What and how will you do work on?
    Spend more time looking forward, not backward.
    Way too often, very good applicants write letters in which they focus solely on what they have already done. The thing is, I already know that from your CV, and it is basically never exactly the topic of the project I’m hiring on. Instead, use the job listing carefully, and perhaps any relevant recent papers from me, and try to think about the future, not the past. You can do this by thinking of a few research ideas you are interested in, and developing them briefly. It’s good to connect this to your past experience, but I am more interested in the future than the past. For example, when applying to a machine translation job:

    👎 “In my Master’s thesis on automatic generation of SQL queries, I have studied the ability of Transformer architectures to learn syntactic structure by mapping their activations to […]. The conclusions of my thesis are that such models can learn structure. I am interested in applying this to machine translation.” — the balance leans too much toward the past. Apply how, why, and has anyone already tried? Show me that you have given your next 4 years some serious thought.
    👍 “In my experience of using automatic translation into Turkish, I have observed many errors for longer, morphologically inflected words. I believe this can be an area where I can use my background in […] to contribute and improve performance. Current work on the topic explores […]. I propose extending this approach with […]. If this works, I would like to explore extending this approach to other languages: this might be challenging, because […].” — Much better! Not a detailed plan, but enough to show you gave it serious thought.

    Do note, importantly:

    • This plan is not a commitment, and does not have to be detailed. Its role is to demonstrate how you think and structure research ideas. Think of it as a conversation starter.
    • This plan should be aligned with the job call: often, the job call is within a funded project that requires work along well defined directions.

    A great letter will basically give me an picture of how our working together and our communication will be like. It’s not about showing off experience or skills: I already know these from your CV. So let your personality show, and try to paint a picture of your motivations and goals.