FAQs - Preparing, Applying and Deciding on Graduate Schools

I received my Ph.D. in 2009 from the School of Computer Science at Carnegie Mellon University. I am currently an Associate Professor of Computer Science at IIIT Delhi. In the last 8 years, I have written Letters of Recommendation(LoR) to 70+ students, and the total number of LoRs that I have written is 350+. List of Universities where my recommendees have got admits are (arranged in the decreasing order of number of admits or accepts, GaTech being the highest): GaTech, CMU, USC, ASU, Cambridge, Princeton, UIUC, Ohio, Rice, UCLA, UMich, NUS, NTU, Stony brook, UF, ETH Zurich, JHU, Lero Research Centre - Ireland, and few others. The period of Sept / Oct - March / April is unnerving for many students who are keen on graduate studies. I am super lucky for working at IIITD, where a lot of very good students go for graduate school (especially Masters). Here are the students who started graduate school in 2017 and 2016 for whom I had written letters. During this period, there are a lot of questions that get posed to me about the admissions process, programs to choose, letters of recommendation, how to decide on which school, etc. For the last couple of years I have been maintaining an internal group level document where we have captured some of these questions and with some of my answers. Given that these will probably be the questions most students applying from India may be asking, I have created this page to share the questions and my responses for the same. Feel free to write to me at pk [AT] iiitd [DOT] ac [DOT] in if you see any questions that you are looking for and you are not able to find an answer here, I will be happy to answer and add it here. This page is mostly intended for the students applying for Masters program in CS / allied areas.


  • What is the appropriate number of grad schools one should apply to? [Go to top]
  • Depends on how much money you have and how much effort you are ready to put in this! Mostly students apply in the range 8-15, 10 is generally a good number to cover all categories. Whatever the number, I suggest distributing these schools in 3 buckets -- 20 - 25% in dream schools (like say Stanford, MIT, CMU), which may be really really hard to get through, but, if you get an admit, you are definitely taking it, no second thoughts; 20 - 25% in safe schools, there is a very high probability that with your application quality, you should get an admit in these schools; and 50 - 60% in probable schools, it is not fully clear whether you will get an admit here. These are the most difficult ones to decide and when you get admits from here, it is hard to make a decision too.

  • How do we go about deciding the safe / probable schools? [Go to top]
  • Parameters that you should look for

    1. Ranking of the institute / program
    2. Fees*
    3. City / Living cost* (this can be a big factor in the US, there is a huge difference in cost of living in different cities in the US)
    4. TAship / RAship* (and other funding opportunities)
    5. Look at profiles of students who got accepts in the past, if there are any from your own college/university, please reach out to them.
    6. GRE cut-offs (not very accurate) but it can help you to get a sense
    7. Talk to seniors/acquaintances studying abroad, I am sure you will use your Facebook or other services mentioned later in the page to ask your questions, including Quora
    8. Look at professors who are working in the area that you would like to specialize and get thesis, etc. Sometimes the school may not be top ranked, but they may have an excellent research group in a particular area.

    * May play a more significant role when deciding on the school.

  • What should be the general structure of a SOP (components to capture in SOP)? [Go to top]

    1. Be truthful (don’t fake it!) most professors reading it are smart to get that. Assume that your audience is a set of people who’ve read multiple statements in their careers.
    2. Focus on the basic idea of why you are interested in studying CS / Information Systems / Human computer interaction / Design / whatever your areas of interests are
    3. Excite the reader about your excitement about CS and related areas
    4. Talk about some courses that you took which awestruck you into CS
    5. Talk about some projects you did which enabled you to figure out what you are excited about
    6. Write it like a story and not like a technical report. Don’t be factually incorrect, or act like your knowledge is superficial. Convince the reader that you know your strengths and weaknesses but don’t lose the narrative by drowning in complex details.
    7. You should also briefly mention about the professor whom you are interested in, not necessarily always working with, but excited about probably taking courses from him, etc. Touching on some of their recent work shows you’ve done your homework.
    8. If there are any research labs in the institute/university that you are applying to which is relevant, you should read about it and make a connection in your SOP. If there are ways you think you can contribute to the lab, briefly mention that. Don’t go overboard!
    9. Keep the SoP in easy-to-read format. The person reading the SoP may not be expert in your area, but you should expect him / her to be intelligent / CS savvy enough to get what you are trying to say.
    10. Please get the SoP reviewed by non-CS people also, this can give you insights into how to pitch it holistically.
    11. SOP should be a well-knit story, it is not necessary for all projects in your CV to be added into the SoP, get only those which are connecting the dots and helpful in making that story.DO NOT make the SoP a simple elaboration of your CV/Portfolio. Try to tell the story that’s not completely captured in CV/Portfolio.
    12. When you review your SoP, ensure you have covered these points:
      1. Who you are (specifically what you want them to see you as)?
      2. Where you’re coming from - what motivates you?
      3. Why do you think you’re the right fit for that institute/university?
      4. How will you contribute to your field after the experience you will gain?
    13. Ensure your SoP has a distinct voice, which is yours. If one can replace your name with someone else's and the SoP still reads the same, then the SoP needs fixing.
    14. Don’t plagiarize the SOP, i.e. do not take something from the internet or your friends, and update it with your details.

  • How much value does contacting a professor from the school carry with respect to your admission process? [Go to top]
  • None. If any, may work in negative most of the times. I strongly think Professors, generally receive a dozen emails everyday from students / others interested in working with them. For a Professor like me, I get at least 10 - 15 emails a week in this context, so you can imagine how it will be for a Professor sitting in CMU or ASU! So, unless you have some strong reason/ connection to the professor, do not write to him / her.

  • BFor a CS program without any intended specialization, should we dig deeper into our research interests or just keep our interests open in the SOP? If we state our research interests clearly, can it in anyway harm our application (for MS in CS)? [Go to top]
  • It should be one level below CS, i.e. it should be at the level of Machine learning / Systems / Security / etc. and not something like Data science for health care from third world countries i.e. too specific. This will help the admissions committee to gauge your application better, but, be open for change, if needed. Strike the balance between focus and fluidity. If you come out as someone who is very rigid, you won’t be considered a good fit.

  • Whom should I choose to be my Recommender / Whom should I take LoRs from? [Go to top]

    1. LoRs play be a BIG role in admissions, so, choose carefully. It is not just the letter that they write matters, it is also the form that they fill on the admission portal that matters. The form generally asks for, intellectual capability, can work in group, can work individually, technical skills in the chosen area, rank the student compared to others being recommended, etc.
    2. A thumb rule for picking the recommender is, he / she should have a lot of good things to say about you, you should have worked with him / her closely, otherwise, the letter will not look strong or not have a lot of good stuff in it. e.g. ‘He/She took a class with me and scored an A’ is not sufficient.
    3. Avoid writing drafts for the LoR, it is ok to send points on what all needs to be captured in the letter, but, not paragraphs / or the entire draft. Each recommender has a different style of writing and you don’t want all your letters to look the same.
    4. Be very clear with your prospective recommender if he/she can write an strong and honest recommendation. Otherwise, both of you are in an awkward situation. A poor recommendation can be disastrous.

  • Should I pick 2 academic and 1 industry / others for LoR or should I pick 1 academic and 2 industry / others? What is the right combination? [Go to top]

    1. Keeping the above response in mind, it is not industry / academia that is going to differentiate, it is the letter itself, a strong and versatile letter is the key. Nothing against industry, but if you can get very strong letters from Academia, I will recommend that. This is a general statement.
    2. Sometimes, it is also ok to select more than 3 recommenders and use them depending on the school / program you are applying to.

  • Should I pick a recommender who is very seasoned / popular / famous, vs. a young faculty or somebody who is not that popular? [Go to top]
  • Keeping the above answers, it is again going to be one who can write a very strong letter. Whoever is ready to do that, you should grab them.

  • Sometimes I("PK") get an email which goes something like this “I got admitted into X university for A program, and Y university for B program, which one do you think is a good place to go?” [Go to top]

    1. The idea for this page was triggered by one such emails. It is hard for anybody to help you make a decision over email.
    2. The decision is very personal, i.e. many personal expectations / attributes play in deciding, so, please remember people like me can help ask some appropriate questions to you which can help you do the rule-of-elimination, but, it is you who has to make the final decision. I am sure you are smart enough to make a good choice. But feel free to chat with someone if you are in a dilemma as to which school to choose from two good schools.
    3. Options mentioned in Qn 2 comes in play here.

  • Does the university keep a track of whether one has applied before or not (moreover, does it matter)? That is, assuming I(student) don’t get an admit from X this year, will it hurt my chances if I apply again next year? [Go to top]
  • Not really for normal applications, if you do anything wrong, they definitely will blacklist you and your university too. Some university applications will ask you whether you’ve applied before and what the result was. You should worry about this problem only in such a scenario.

  • Is it OK to provide a link in your SOP to some of the work I've done such as a publication? [Go to top]
  • It is ok, but keep it minimal.

  • Should / can we write to professors for TAship / RAship before going to campus? [Go to top]
  • As explained above, professors get a lot of emails, so you should avoid. Also, if you take the admission and using your university ID, etc. @andrew.cmu.edu ID if you send an email to the professor, it is going to be received better than a gmail or a IIITD email ID. I am strong promoter of the idea that you take a course (or an independent project) of the faculty whom you want to RA / TA, get an A in the course, and then reach out to him / her asking for RA / TAship, you are more likely to get it.

  • Should we apply / accept generic programs like MS in CS or specialized programs like MS in Data Science or MS in Security or MS in HCI? [Go to top]

    1. This is very subjective, but, majority of the times in such a situation, students have chosen the option of going for generic MS in CS programs, even though the school ranking is low, etc. Rationale being, taking specialized courses / programs may restrict job opportunities. There are students for whom this choice has not made a difference.
    2. But if you are really really sure you want a career only in HCI or some such specialised area go for it. Another aspect of this decision is what you get to learn from the specific course/program. If you’re putting in the time, money and effort to go do a Masters you should make sure that you’re not repeating the same stuff you did in your undergrad. Choose a program based on how much you’ll learn from it and how much it will help you to achieve your goals.

  • What are the scholarship opportunities for students going from India? [Go to top]

    1. JN Tata Scholarship
    2. TOFEL scholarship
    3. KC Mahindra Scholarship
    4. Narotam Sekhsaria Scholarship
    5. Rotary Foundation Global Scholarship
    6. Some more relevant ones can be found here

  • What are the online services / websites that help in answering questions regarding to admissions? [Go to top]
  • Some services that students from India use are

    1. The Grad Cafe
    2. Yocket
    3. CS Rankings
    4. How to apply for Ph.D. programs?
    5. Details about graduate school life and expectations
    6. Graduate schools ranking: computer science
Thanks to Megha Arora, Vedant Swain, Ashwin Rajadesingan, Rohan Katyal, Amod Agrawal and Yashovardhan Sharma for inputs in various parts of the article.