For Prospective PhD Students

I am looking for 1 or 2 new PhD students to join us in the fall of 2025. The application deadline this year is December 15. Here is how to apply to our program.

Our main topics of interest at the moment are human-AI interaction (broadly construed) and accessible technologies for people with motor impairments.

What are the most important factors you look for in an application?

Broadly, I try to assess if a candidate can think like a researcher, has the necessary academic preparation, and whether they have the curiosity, grit, and perseverance to take on risky problems, and to not give up in the face of inevitable setbacks.

Another general point: I evaluate each candidate with respect to the opportunities they had access to. For example, I expect research experience only from students who went to schools that made research options readily available to undergraduates.

Here is some specific advice I give to students I work with:

  • Use your Statement of Purpose to demonstrate that you can think conceptually. If you have had research experience, make sure that you can communicate the scientific purpose of the project you were part of and how your own work contributed to the big picture. Don’t just write “I implemented 4 features using these 7 libraries”.
  • In your Statement of Purpose, you also need to communicate what you are interested in. This is a little tricky. On the one hand, it really helps if you can include some specific examples of problems that fascinate you. On the other hand, you do not want to create an impression that the example problems are the only ones you are interested in. An effective strategy I’ve seen some applicants use is to say something like “I am broadly interested in problems [that have the following characteristics]. For example, …”. Also, remember that PhD research is about creating generalizable knowledge that does not yet exist (while most of undergraduate courses are about applying existing knowledge to solve specific problems). While you are not expected to translate your ideas into the technical vocabulary of your intended field, try to identify general challenges rather than specific products or solutions you’d like to create.
  • Make sure that the majority of your letters are from researchers (either at a university or at a research lab). Letters from non-researchers rarely contain information helpful for evaluating PhD candidates.
  • Grades are not super important (to me). A mix of As and Bs is fine. Lots of Cs, however, is a red flag — make sure to explain those (e.g., if you had a rough semester or two). The first thing I check on a transcript is what courses a person took. Did they take relevant courses? Did they take challenging courses? I’d rather see some Bs in challenging courses than As in easy ones.

Can international students do a PhD at Harvard?

Yes! Here are some important things you need to consider. First, the application process takes nearly a year. Like most graduate programs in the US, Harvard requires applicants to take TOEFL exam (for non-native English speakers). Our applications are typically due in December or January, and the results are announced in March or April for programs starting in September. Second, the cost. In Computer Science, you should expect to receive financial assistance that will cover the tuition and that will also provide a stipend sufficient to maintain a frugal but perfectly reasonable lifestyle. This financial assistance can come from a fellowship, a research assistantship, or a teaching assistantship.

Can I email you about the admissions process?

I generally do not respond to admissions-related inquiries. First, I want to give as much time as I can to people who are already working with me. Second, I don’t think it would be fair. People who have the confidence to email prospective advisors tend to be already relatively privileged. By providing them with personalized application advice, I would only be increasing inequities in our higher education system.