PhD application is a confusing, frustrating, and unique journey filled with many variables. There is no telling what might happen. After getting through to the other side (although the real challenge is completely ahead of me, hehe), I thought I should document it all somewhere; maybe it could be of use to someone else. But a disclaimer before you read any more: what worked for me may not work for you at all. There are too many factors at play here; no 'one formula for PhD admission' exists. This is solely based on my own experience. These are just my suggestions which may not be right for you.
I wrote this to help anyone who is thinking about applying to a PhD program. I want to help as much as I can, but as you will soon find out, PhD is insane and I honestly do not have time to reply to texts, even though I really want to. So, if you have any questions to ask me directly, please go through this document thoroughly first and be absolutely sure that the answer cannot be found here. Also, for a lot of this, you have to use your own intuition like I did. So, only ask me questions if you absolutely need to, and please forgive me if I fail to reply (even if you can't forgive me now, you definitely will after you start your own PhD haha)
Mushtari Sadia
Lecturer, Dept of CSE, BRAC University
Adjunct Lecturer, Dept of CSE, BUET (July 2023 Term)
PhD Applicant, Fall 2024
Field: Systems/Security (CS)
Undergrad CGPA: 3.84 (CSE, BUET)
GRE - Verbal: 161/170, Quant: 166/170, Analytical: 4/6
TOEFL - 115/120 (Reading: 29/30, Listening: 30/30, Speaking: 30/30, Writing: 26/30)
Number of publications (during application): None
Application Status:
- University of Michigan - Ann Arbor (Interviewed, Accepted + Funding)
- Purdue University (Interviewed, Accepted + Funding)
- University of Massachusetts-Amherst (Interviewed, Accepted + Funding + Fellowship Nomination)
- University of Virginia (Interviewed, Accepted + Funding + Fellowship)
- Virginia Tech (Interviewed, Accepted + Funding)
- Stony Brook University (Accepted + Funding)
- Penn State (Interviewed, Rejected)
- Yale University (Interviewed, Rejected)
- Northeastern University (Interviewed, Rejected)
- University of Washington (Rejected)
- CMU (Rejected)
- UC Berkeley (Rejected)
- UIUC (Rejected)
- Johns Hopkins (Rejected)
- Maryland College Park (Rejected)
- UC San Diego (Rejected)
More about me can be found on my website: https://mushtari-sadia.github.io
There are two semesters when a PhD program begins, Fall (August) and Spring (January). Usually most programs admit students with financial aid during Fall, which is why most people target the Fall semester. The application deadlines for Fall semesters are usually around December 15 of the previous year. For Spring semesters, the deadlines are around August 15. Admitted PhD students usually receive full tuition waiver + monthly stipend in the form of either a research assistantship (RA) under a professor, or a teaching assistantship (TA). If a university really wants a student, they may offer them extra fellowships on top or instead of the RA/TA funding.
Each application to a university will require an application fee. Typically, people apply to 8-12 universities. The application fee for East Coast schools can range from $60 to $100 USD, while West Coast schools may ask for a whopping $155 USD per application. Sending GRE score to each university also costs $35 USD, while sending TOEFL score costs $25 USD. Many universities offer application fee waivers based on proof of financial hardship or attendance at selected programs/conferences. I received many fee waivers because of my attendance at the GHC conference (only for women). So if you're a female CS undergrad student, take a look at here.
To apply to a PhD program, you’ll need the following components as part of your application:
Each and every component plays an important part in your application. The strength in one may make up for the weakness in others. I’ll talk about each component as it comes up. I have tried to organise everything as a timeline, so you know what you should be doing and when.
So, what will be the most important factor in your application? A great CGPA? Excellent GRE, TOEFL scores? Publications? While these are definitely awesome things to have, they cannot guarantee your admission alone.
The one thing that can increase your chances of admission the most is, alignment of interest and experience with a professor. That is, having a good amount of research experience in a specific area, and finding a professor who is interested in the same area (ofcourse, the professor should also be looking to hire new students). The "good amount" can be different for different fields. For more competitive fields like ML, CV, NLP, it can mean having published first-authored papers in top conferences. In my case, I had two major research projects in security. None of them were published at the time of my application.
To obtain said research experience, you should look for projects to be involved in, in your field of interest. A good starting point is your undergrad thesis.
It is almost impossible to completely be sure that you will want to do a PhD in the same field before choosing your thesis. Heck, I didn’t even understand half the words of the topic when I chose mine. However, your thesis is a research project you will spend a year on. It doesn’t matter if you can accomplish something new or achieve state of the art results with your thesis. Most undergrad thesis are, well: stupid. It’s supposed to be. But your thesis makes you familiar with the recent works in a field by literature reviews, work with a specific problem for a long time and identify future directions. These things make your undergrad thesis a significant strength in your PhD application. Again, and I cannot stress this enough: professors mainly care about your research experience that matches their own interests. If you eventually decide to change fields, your undergrad thesis will no longer be a strong point in your application. So, try to choose your undergrad thesis wisely. One idea is: You can visit the websites of top universities in the US, find the faculty members in your department of interest, check out a few of their papers to just get an idea of what they have been working on recently. It's okay if you don't understand the topics. Just know the names of the topics. Look for keywords listed in their papers. This is just an idea - I did not do this myself. I had zero clue, and chose my thesis for two reasons, 1) because I was interested in systems/security, and 2) because it was a collaboration project. I thought I could get 2 LoRs from one project, which I did.
To make your application strong, you need as much relevant research experience as possible. The key word here is - ‘relevant’. For example, if you’re applying to security, your published paper in bioinfo will not matter very much. So, look for research projects in your field. Here, one of the application components should be mentioned:
The first component you should work towards is the letters of recommendations, or LoRs. To get LoRs, you need to be involved in research with at least 3 researchers.
Usually a lot of people just get LoRs from their thesis supervisor + two course teachers who don’t know them very well. However, the ideal and most effective LoRs come from people who are reputable researchers in your field + have direct research experience with you + on projects that are relevant to the field you are applying to + and ofcourse, have good things to say about you.
In my case, I am grateful to receive LoRs from the following people:
1. My undergrad thesis supervisor (One LoR must be from your undergrad school, most preferably your thesis supervisor, if you’re applying to the same field as your thesis.)
2. My undergrad thesis collaborator (Research Officer (Cybersecurity), National Research Council, Canada.)
3. Collaborator in another research project in security (Visiting Assistant Professor, MIT)
4. One of my course teachers (It is wise to manage four recommenders in case you need a backup)
A LoR from a well-known figure within the academic community (in your field) in the US, can be everything. In my case, multiple professors directly mentioned to me that they were impressed by the third LoR (from MIT), as they were familiar with his work. I was connected with him through the Fatima Fellowship program.
So, try to be involved in different research projects in and outside of your school, collaborating with different people. Again, ideally, these projects should be directly related to the field that you are applying to.
This is when I took my TOEFL exam. It’s not necessary to take it so soon, many people took their TOEFL in September/October. This is basically a compulsory English proficiency test. It does not require too much prep, so it’s good to take it out of the way sooner than later.
When registering for the TOEFL, you have the option to select four universities to which you can send your scores for free. Try to pick four safe choices- places you know that your chances of applying are high.
Some universities waive the TOEFL/IELTS requirement if you can upload a document containing proof that the official language of instruction of your undergraduate institution was English. BUET provides a certificate called "Medium of Instruction", which you can obtain by applying in the "Other Certificates" section in BIIS. While this can save you some money (since you have to send your GRE/TOEFL scores to each institution incurring separate charges), I do not recommend skipping your TOEFL test by relying solely on this certificate.
Create a CV and a personal website highlighting your research experience and accomplishments. For reference:
This is also the time when you should start exploring and listing professors, universities.
These websites may help:
2. csrankings
Go through the faculty pages of individual websites of the universities to find professors in your field. Look for newly joined assistant professors, they are most likely to hire new PhD students. Comparatively more senior professors also look for new students sometimes. If they are looking for new students, they usually put it up on their websites. Save the links in a google sheet neatly.
Note the department of the professor when you're listing them. For example: I mailed two professors at Penn State, without noticing that one is affiliated with the CS department, and another with the IST department. Both departments conduct research in security, and I could apply to only one department. You will have to write the names of at least three professors with whom you want to work with in your SOP, and in most cases, all three professors will be sent your application. If you apply to, suppose the CS dept, and write the name of a professor who is in the IST dept, you are wasting one name.
This is the time when you should start emailing professors. This is not a part of the formal application process, so some people skip this step, and directly apply to the program in December. This is very risky, as high-ranked universities get thousands of applications each year, and there is very little chance that they will pay attention to your application/your application will stand out in the central review process. To get a professor’s attention, you must reach out to them early and preferably before others.
For example, I only got into universities where I previously reached out to a professor and got a positive response. I received straight rejections from universities where no professor replied to me/I didn’t reach out to anyone.
Carefully craft an email to a professor before sending. Go through their website beforehand, some professors put up specific instructions on how you should mail them. Get an overall idea about their interests.
People usually say you should read a bunch of papers of a professor and come up with something smart and analytical to say about those, before emailing them. I did not find this suggestion practical. First of all, I did not find it feasible to read years worth of research and come up with a smart question about them in a few hours. Secondly, do professors really want you to ramble about their past work? Those are already completed research projects.
They want NEW ideas from you, they want you to work on NEW projects. They want to know about YOU! So focus most on how you can present your research most effectively. Ofcourse, that doesn’t mean they don’t want you to prove your interest in working with them. So, go to their google scholar page and include two of their recent papers that are most aligned with your own interests in the mail. You can use tools like pdf.ai to quickly read the papers (or just read the abstract + last paragraph of introduction + conclusion). You can write a few sentences about the papers, but I would not put too much effort on this part.
Here is a template mail:
Here’s another:
Choose the ‘schedule send’ option in gmail to send all the emails on Monday evening, which is Monday morning in the US. You can also send the emails on Tuesdays to Thursdays, but do not send them on weekends.
Do not get frustrated if you don’t receive responses. Keep sending emails. I mailed about 36 professors, received around 12 replies, and finally got 6 offers. One of my friends sent over 40 emails and got like two responses, and she got into a top university as well. There is no rule for this. Most professors will NOT reply to your email, and that’s okay.
If you do get a reply, there are three type of responses:
The best time to mail professors is August to mid October, but I’ve mailed professors until late November. So, keep mailing. The professors from whom I eventually received admission offers, all responded positively to my emails.
If you get an interview from an email, go to section - January 2024.
This is when I gave my GRE exam. I regretted taking the GRE, because most universities don’t even consider GRE in their admission process anymore. Professors do not care about GRE at all. If you have a decent CGPA, I strongly suggest not going through this unnecessary hassle. For me, it was a complete waste of time and money.
However, if you do decide to take the GRE to be safe, one helpful tip: look up the institution code for GRE of four universities that you are planning to apply to before going to the exam venue. Also, check if the university actually considers GRE in their admission process or not. After taking the exam, you will have the option to add the codes of four universities where you can send your scores for free. Like TOEFL, they don't let you do this while registering for the exam.
Time to start finalising universities. Go through the websites of the universities and start listing required application materials. Many universities offer application fee waivers based on some criteria. I received fee waivers from 8 universities by showing financial hardship or attendance in the GHC conference. (Read this to learn more about fee waivers).
I also listed everything on a google sheet which came in very handy.
Also, you can start writing your SOP during this period. Some universities ask for a personal statement along with your SOP. The main difference between the two is, SOP should be all about your research, and personal statement can be about the challenges you've faced in life/other accomplishments. Each university looks for different things in the essays, and they mention what they want in their websites. Here are some helpful links for SOP writing:
While writing: Keep a specific purpose in mind for each paragraph, and make each sentence mean something. Do not write any irrelevant/unnecessary sentences. For each of your research experience, remember to specifically include your personal contributions + outcome of the project + possible future directions. Do not use ChatGPT or AI tools for writing - just use them for proofreading.
There are GAAP programs which will arrange SOP reviews for you. You can also ask your seniors to review your SOP for you.
Here is a version of my SOP:
My personal statement: (this is required by some universities, but I do not think this carries much weight for PhD applicants. This maybe more important for undergrads. However, the specific guidelines for writing personal statements are usually provided on the websites of universities)
You should also email your recommenders during this period to formally ask them if they would provide LoRs for you in your upcoming applications. They may ask you to send them a draft. Prepare the drafts too. While writing LoR drafts: try to write concrete facts about your contribution in the project with the recommender. Avoid using billions of adjectives.
My LoR drafts:
You should start creating application accounts in admission portals. Gradually prepare and complete all required application materials. Remember to add your recommenders as early as possible so that they have plenty of time to send your letters. You can prepare them a sheet like this.
Submit applications. You can email professors to let them know that you have submitted your application.
Template:
Dear Professor,
I submitted my application to the CS, PhD program at UVA (Reference No: 545517271), and included your name as a potential advisor. I hope you will consider me for any available positions in your lab if I'm admitted.
Regards,
Mushtari Sadia
Adjunct Lecturer
Dept of CSE, BUET.
Website: https://mushtari-sadia.github.io/
Professors should reach out to you during this period for interviews. I got a total of 9 interviews, 3 during the mailing period (before submitting applications), and the other 6 in December-January.
Create a portfolio slide. For reference: my portfolio slide. The slide should mainly include your major research projects. Include a summary page for each project first, then pages that delve into more details.
Get familiar with the professor's research. Again, you don't need to fully understand or memorize any of their work. Just skim three or four of their papers, go through their website to get an understanding of the domain of their work. Before you get an admission offer, the interview is about you, not them. So you need to be prepared to talk about you, not about them.
Prepare yourself to talk about your research in accordance with the slide, revise your work so you can answer questions from them. It's understandable if you don't know the professor's work, but it is not acceptable if you cannot answer questions from your own work. So focus on being prepared on that area. Organize your thoughts. Also, it will be very impressive if you can talk about some problems you want to work on in future. For example, after talking about my thesis, I told the professors - 'I faced $these$ limitations during my thesis, and saw that there is a lack of research/solutions in $these$ specific problems. If $this$ could be done, $this$ would solve $that$ problem. I want to work on that in future'. The professors will not hold you to actually work on those problems in future. But it demonstrates your capability of critical thinking.
You should start receiving offers during this period. This is a very difficult time. Try to be patient.
I'm not sure if I want to do a PhD. I can't decide. How can I be sure?
- I get this question a lot, and honestly, I can't help with this very much. I myself wasn't always sure, but eventually I knew I liked being in school, I wanted to do a thing called research where you get to work on a different range of challenging problems, and I wanted to eventually become an expert on something. You know when you know. If you aren't sure PhD is for you, you can take some time before you decide. Do not rush this decision. However, remember that it's hard to be a 100% sure about anything.
What do I write as a reply to this email from a professor?
- I feel like such a hypocrite saying this, because I myself bugged my seniors asking this question a lot (shoutout to Adiba apu, love you forever and ever), but, you should actually use your own intuition to write these replies. In most cases, just say thank you and move on. No need to overthink that much. Again, there is no exact format in doing these. Do whatever feels right to you.
Would you review my mail/SOP/resume?
- If I have the time, absolutely! This is the correct thing to ask me about directly. Don't ask me/anyone else what to write, write something first and then you should ask someone to review. Again, if I have the time, I'd be happy to! Also, there are many resources I have shared in the October 2023 section for writing materials. Check those out!