Ryan K Cosner

Ryan K Cosner

Caltech PhD student studying safe control and learning for robotics.

PhD Application Insight and Advice

Date: 11/5/2024.

My Application Materials

I don’t think that these were the final drafts of these statements, but I lost access to my old Berkeley email address, so I can’t find the old files anymore:

General Application Advice

In general applications to PhD programs are more similar to job applications than they are to undergrad applications or even other grad applications (i.e., med school, law school, business school). Critically in this process, you are applying for a position where they should be paying you, which changes the calculus. The PhD model is that, they generally want to accept every student that they can, but they are limited by funding.

Personal Statement (aka Statement of Purpose)

The personal statement is like a sales pitch where you are selling yourself as a research employee. This should read more like a cover letter for a job, than a self-reflective undergraduate application essay.

It’s normal, but not always true for you to be accepted into the school by a single lab with the understanding that you will work with that advisor. This is very different from undergraduate applications where you accepted to the school more generally. Because of this, you want to tailor your application so that specific people/professors/labs will be interested in your work and skillset. This doesn’t mean that you have to choose exactly 1 professor that you want to work with, but find a handful (~3-6) in your field that you could see yourself working for.

As an aside, if you cannot find at least 3 potential advisors at a school, then you should not go there. It is very possible that 1 advisor might not have funding, another might be a bad fit, and then you’re left with no one to support your studies. So you should only go to a school if there are at least 3 professors that you could see yourself working with.

In general, putting a professors name on an application is very low commitment. Basically, you’re signaling that you want those professors to review your application. This is similar to the peer review process where we mention keywords associated with our field to make sure that we get reviewers with the right expertise. You want to make sure that your application gets in the hands of the right professors who are the most likely to hire you.

Your personal statemnet should contain the following information:

  • Intro
    • I have these interests
  • Body
    • I have these skills
    • I have these experiences
    • I can make these specific contributions
  • Conclusion
    • Here’s why you should hire me

Since advisors are hiring you for your research skills, you should focus on your prior research contributions. Talk about how you contributed to projects in a concrete way and how your skill set will help them/you complete their/your research goals. Prior publications are also really really helpful, especially if they’re in relevant venues. In my fields of robotics, control, and machine learning these would include (but are definitely not limited to): ICRA (robotics conference), IROS (robotics conference), ACC (controls conference), CDC (controls conference), RSS (robotics conference), CoRL (robotics + learning conference), L4DC (learning + control conference), NeurIPS (learning conference), ICLR (learning conference), CVPR (computer vision conference) and associated journals like TRO (robotics), TAC (controls), Automatica (controls), Science: Robotics (robotics), etc.

Personal History Statement

Some schools used to have this secondary statement which was similar to an undergraduate application essay and a faculty diversity statement. You can use this essay to introduce yourself more personally. I really value the content of these essays, but that is my personal belief, some professors say that these are significantly less important than the statement of purpose.

GRE and Testing

A lot of schools are phasing out the GRE which is probably good. It was a really expensive test that wasn’t really that useful anyways. I’ve heard that the GRE couldn’t ever really help you, it was just a standard that you needed to reach.

And for engineering you only need to take the general GRE at most, the subject tests aren’t important.

Funding

Since PhD programs should pay for your education, funding is a huge factor in the admissions process. A cynical perspective on academia is that graduate students are poorly paid research scientists (I believe this is only partialy true, grad school can be a phenomenal learning experience as long as it is framed that way). Thus, if an advisor can have research scientists working for them for free, they are very likely to accept them.

If you bring your own funding, the paradigm shifts, and it can be more of a traditional advisor-advisee relationship and less of a boss-employee relationship since you’re not dependent on your advisor’s grant funding.

Off the top of my head, there are a handful of general fellowships which you can take with you to whichever school you choose to attend:

  • NSF
  • NASA
  • NDSEG
  • Ford Foundation
  • Hertz Foundation

Some schools have school-specific fellowships:

  • Stanford’s Knight-Hennesey

Getting Admitted

There are various versions of “getting admitted”, but they generally fall into 3 categories:

  1. Admitted to a PhD program with funding: This is what you want. This means that the school/advisor has resources and is willing to invest those resources in you. Generally this funding lasts a year to a couple of years and then you’ll be funded off of other things like grants or industry collaborations.
  2. Admitted to a master’s program with potential for PhD: This is kind of a middle ground. Some of the more traditional, prestigious programs do this. Be careful. If the master’s is funded, then it’s not necessarily a bad sign, but check to see how hard it is to convert from a master’s student to a PhD student.
  3. Admitted with no funding: This is generally bad. You should not enroll in a PhD program that you have to pay for. And more generally, you should not be paying to do research.

Timelines:

Fellowship applications are generally due in October and can be a great way to get your ideas down on paper and refine your research interests even if you don’t get any of them. For example, I applied to NSF and didn’t get it, but it was super useful to help me refine my research interests and to make sure that my letter writers got my letters in for the October deadline so that they were ready for the school deadlines in November/December.

Choosing Schools

Public vs. Private:

  • Public Schools tend to have more teaching money, meaning that you work more, get more teaching experience and have more collaboration
  • Private schools generally have money coming from endowments, so there are more scholarships and free money floating around, but there are less undergrads so there are less teaching opportunities and more individualism

General Recommendation:

Some advice that I got that I found useful:

  • This is going to be one of the first times where people might outright disagree with the life path, so it’s important to take initiative, stand proud, know yourself and make decisions. For example, my dad never understood why I didn’t just drop out with a masters to earn more money sooner.
  • When choosing a school, you are actually choosing a lab. Yes, you should like the research directions that the lab has, but more importantly, you should like the advisor and the other grad students. Grad school is about intellectual discovery, so it is very likely that you’ll find new reseach directions that you’ve never heard of before. And that’s totally great! You just want to make sure that you have a supportive, caring, empathetic, and compassionate environment to make those discovers. Make sure that you choose a place where you feel comfortable taking academic risks, like exploring new ideas and sharing thoughts even before you’ve had a chance to fully polish them.