William Chen

William Chen

Graduate Student at CMU LTI

Grad School Apps

This page contains an archive of my past grad school application materials along with some general thoughts about the process, generally targeted to students interested in speech and audio processing in the USA. I participated in two admissions cycles: F22-S23 (Master’s and PhD) and F24-S25 (PhD only).

F22-S23 Cycle

CV

SOP

I applied to 6 schools: CMU, Stanford, University of Washington, Johns Hopkins University, New York University, and University of Colorado Boulder.

I applied to CMU and Stanford’s masters programs, since the former paid a stipend (for MLT) and the latter was doable online. I also applied to CMU’s LTI PhD program. For all other schools, I applied to PhD only.

I got interviews at JHU (CLSP) and Boulder, but was ultimately rejected from both. The CMU MLT was my only acceptance. I got rejected from all other schools without an interview. Note that in this stage of my research career, I was more interested in language processing (both speech and text) for low-resource languages, rather than broader speech processing.

F24-S25 Cycle

CV

SOP

I applied to 7 schools: CMU, MIT, University of Washington, Princeton, Stanford, Georgia Tech, and Johns Hopkins University.

I was rejected from Princeton and Stanford without an interview. I got offers from the other schools with these PIs:

Georgia Tech was confusing. The PI I listed in my SOP was retiring (although I did not know this at the time.), but I got an acceptance without an interview. On my visit day, they had me talk to faculty that appeared randomly assigned - the topics included nuclear engineering, supply chain optimization, and chip design. Everyone else talked to someone relevant to their research. Celin Lin also later reached out to me separately. I would definitely recommend students to reach out to faculty to check if they are hiring. Don’t be like me and waste application money on several schools where the PI of interest was retiring.

There are several other faculty that do wonderful research on speech/audio/music. Some (but definitely not all) include:

In my opinion, research match and taste is one of the most important factors when selecting a PI. Don’t pretend you are interested in something for the sake of getting an offer letter.For me personally, its hard for me to stay motivated if I am not interested in the topic. A PhD takes a long time - spending 4-7 years working on something you don’t like will not leave you happy. I don’t think this is something worth sacrificing for school prestige.

Interviews

Interviews are very school and PI dependent. This includes the scheduling too. If your friend got an interview from Professor foo, don’t panic if Professor bar hasn’t reached out to you yet (although if it has been a few weeks and people are getting acceptances, you are probably out of luck). Some places (or PIs) won’t even interview you before sending an acceptance letter, although I believe that is becoming increasingly more rare.

I found that the interview contents can change depending on how “senior” of a candidate you are considered. People who have more projects and papers published get asked higher-level questions, such as gauging experience, topic match, and research taste. More “junior” candidates are more likely to get quiz-like questions, like “explain ABC to me” or “how does XYZ work”. Of course, your mileage may vary - this is always PI-dependent.

Acceptances and Visit Days

Acceptance notifications will usually come out around late January to late February, although this can always change.

Visit days are really fun. I would definitely recommend going to them if you can. I made many friends during the visit days, with both accepted and current students, and some have led to research collaborations down the line. Note that some schools will put you in your own hotel room, others will give you a roommate (sometimes without telling you!). Definitely try to talk to current students of the PI. Unlike the PI, they do not have a vested interest in you accepting the offer. Their advice will usually be the most relevant - they were once in your shoes. You can also ask them about non-research things, like the stipend, where to live, etc.

Many students like to ask their prospective PI’s about potential co-advising, because they have multiple professors they want to work with. Frankly, I wouldn’t be surprised if faculty are really annoyed with this question. Before you try to be co-advised, ask yourself a few questions. Is there a common interest between these two professors? Do their working styles seem to match? What will you do if they give you conflicting directions/suggestions? Having two advisors means you have more research suggestions, but also more bosses that you need to listen to.