pterosaur

@climate_doge - I didn't get terribly specific in my sections, either. It's not a research proposal like the GRFP. Also, remember your audience here: it's being written for people who know computational science but don't necessarily know your field. So that means not a ton of details (you won't have space), but pointing out important connections between your field/interests/research and high performance computing. For my research, that didn't mean supercomputers, exactly, but the power of parallelization for simulating neural systems in robots.

Just ran across this now, so I'm not sure if you'll see it, OP. In the US, I don't think there's a big difference in the way that a PhD from a typical UK program would be perceived compared to one through a CDT. First, because most people in the US wouldn't know the difference or know what a CDT is. It seems like a lot of the CDTs are essentially taking the more American approach of combining the masters and PhD and putting them into one program. (The 4 vs 5 years then comes mostly from the fact that the UK does 1 year masters compared to the more typical 2 years in the US.) I can't speak to these anecdotes, but in general PhDs from Oxford, Cambridge, and Imperial would be very well regarded in the US. Given my knowledge of the CDT program, I might see it as a tiny boost over the regular UK PhD because there would be a more structured/consistent background knowledge I could expect out of those completing the program, but it likely wouldn't be a determining factor in anything. I'm from the US, did an MRes at Imperial,\ and am now at Harvard SEAS for my PhD. For what it's worth, I got into everywhere I applied in the US but was summarily rejected from an Oxford CDT program without even getting to the interview stage (they interview nearly 70% of the applicants...).

I'm freeeeee* * aka I turned in my final project last night so now I have time to do actual research work.

I guess a bunch of us hang out here to procrastinate. But maybe someone will run across the thread and give it a shot. I only heard about it from 1 friend, otherwise I never would have known to apply.

@marmle: it is possible to apply successfully without already being in discussion with a possible advisor -- I did. I spent a lot of time looking at the course listings and CSGF website to figure out what courses fit my research goals and the purpose of the CSGF, and had my undergrad research advisor (in a rather different research area) sign off on it. Also remember that you will not necessarily end up going to the university in your application (either due to not being accepted or what turns out to be a better fit), and they're aware of that, too. In trying to tailor my applications to the different fellowships I applied to, I ended up listing a different school for the CSGF and GRFP because I actually proposed rather different research for each of them.

I'm going to put in a plug here for the DOE CSGF: https://www.krellinst.org/csgf/how-apply. It's not as well known as the NSF GRFP or NDSEG, but it has some really great perks. It pays a stipend of $36,000/year and full tuition/fees/insurance charged by the university, plus extra money for funding research and travel expenses. You also get to do a 12-week practicum at any of the national labs, and there's an annual conference with fellows, alumni, and DOE national lab reps. It's a fellowship for PhD students (either entering or in their first year) doing research related to computational science. This doesn't mean computer science; it includes students in fields from biology to physics interested in taking a computational approach to research, especially with a focus on applying high performance computing. I'm a first year fellow and really pleased with the program. I'm hoping to get the word out and encourage more people to apply. Bonus: the deadline isn't until January 18, so you've still got plenty of time. And I'd be happy to answer any questions!

If you're looking for what you can do to improve your chances for this application cycle, there's not much time to change things at this point. The main thing you have left to do is nail your SOP. Based on what you've written, you seem like a strong applicant. However, you do only mention MIT. Is this the only school you're applying to? Remember that there's a bit of a crapshoot in applications no matter how good you are, so be sure to give yourself options. Also remember that even highly qualified applicants can be rejected if the ad com doesn't feel there's a good fit between the applicant's and program's research interests.

I don't know ML specifically for what are the top schools, but I think you have a very solid application. You have a strong quantitative background and research experience in CS. Switching from math to CS shouldn't be a problem for you.

It's not very helpful to say, but it can vary a lot by school. They may be slow because of the overload of admissions season, or they may release decisions in bulk. It's hard to say.

At this stage, you could have sections for; education grants/fellowships/awards research experience teaching experience posters/presentations (even a departmental poster presentation is worth mentioning here) other experience (community service, internship in industry, etc) relevant coursework skills (lab/computer skills, whatever is relevant to your field) interests (I had a single line about my involvement in band, curling, and interest in graphic design) But it will seem pretty empty at this point relative to an experienced grad student.

The program you've accepted is expecting you to come and will take into account this "filled" spot in their admissions decisions. It's not fair to them to hold this acceptance while applying to other places. That's my interpretation, at least. You can also contact the university directly.

Harvard div is a really interesting mixture of religious beliefs, including athiesm. Being a Confucian should not be an issue for them.

Not totally academic, but... this election. I'm guessing it hit a lot of people hard. We had a group meeting scheduled for today, and my advisor brought cupcakes and kleenex. We needed both. She was crying and I was tearing up too. We abandoned the planned presentation and instead talked for an hour and a half about what we can do as a group to make the world a better place and improve society and the world, not just focusing on our research. It was very passionate, and it looks like we're really going to expand our outreach efforts. I'm really proud to be part of this group and I'm so lucky to have the advisor that I do.

@Pink Fuzzy Bunny - I had an exam that went almost the exact same way... I was working/studying hard all semester, aced the first problem set, got through the the previous years' exams fine. The profs said the exam would be based on the material from the lecture notes and exercises, and there would even be questions from the exercises on the exam. All of these things were lies. It was a disaster. I've never left questions blank on an exam before (because it was also one of those where parts b,c,d, and e build on part a, so if you don't get the first part you're screwed). The last time I even got a B in a class was 7th grade gym class. I ended up deciding to drop the class, even though it still would have been technically feasible to get the required B- in the class; given how big of a time sink it was (I had no time for research) and the unhealthy levels of stress it was inducing, I decided that was best for my mental health. My advisor is also totally cool with that, which is a relief. I feel 300% better since dropping it last Monday. I'm doing way better than one of my friends in my cohort who is losing patches of hair from stress. But I also found out I got Distinction on my masters, so it's not all bad!

I'm from the US and back in the US for my PhD, but I got my masters in the UK and I'm not sure how to list in on my CV. It seems like in the UK they'd just say "MRes(Dist) in Something" but I don't know if Americans would know what the abbreviations are for. Would I be better off spelling out "Master of Research in Something, with Distinction," which seems kind of superfluously wordy.

Someone in my cohort is seriously considering getting a Harvard baby onesie for her cat and walking it through the Yard just to see what reactions she gets from tourists. So, cats can be a source of endless entertainment in grad school.

I'm really excited by the research I'll be working on... but I don't have time to work on it! I'm taking 2 classes, and they're currently eating all of my time. One requires close to 15 hours a week just for the readings, the other has weekly ungraded exercises (which you need to do if you want to do well on the exams) which can be 14 pages long, and they both have incredibly long problem sets due next Thursday. Arrgh.

@Pink Fuzzy Bunny - Interestingly, the founder of Piazza (a woman herself) first created Piazza as a tool for CS classes, hoping to get more women involved who were intimidated by their male classmates and therefore didn't speak up in class. It was only once she added the anonymous feature that the female students started posting as much as the male students.

Are comps just another terms for quals? I have mine the end of 2nd year, but they're pretty different than most universities do for STEM, I think. We don't have a comprehensive written exam or interrogation. It's a presentation of the research we've done in the first 2 years basically to prove that we can do good research worthy of a thesis. It's apparently pretty low stress. But it does mean that we have to demonstrate knowledge by taking a bunch more classes (10, compared to 6 at MIT for example). I think I'm OK with that, except that I'm currently already in minor panic mode about how this machine learning class is going to go. Also, I finally finished unpacking all the little odds and ends today as a form of procrastinating on a final read-through/edit of my masters thesis. (Now that I've gotten to it, I sound way less idiotic than I thought I would.)

@janetjanejune - Thanks for the recommendation - I'll have to check that out! I'm also now in the office at nearly 8 pm on Sunday because of course I still haven't finished my bloody master's thesis that's due in eight days. I kept saying I'd get it done before I left the UK, but of course I didn't. It's really coming back to bite me now that I also have a crap ton of readings for class. Wonderful, math-heavy, algorithm-filled, 25-page, computer science theory papers...

In addition to the globs of undergrads all over campus,I have another obstacle I didn't expect: tourists. They're all over Harvard yard, in an often creepy way. As in, the people on the first floor of my dorm said they've had tourists trying to take pictures through their windows so they know what Harvard res halls are like. I feel like I expend a non-insignificant effort trying to make clear that I am not a tourist.

I'm also building a computer! My department gave me a $1350 budget, and I managed to spend $1347 of it. I also managed to squeeze in a Gateron brown mechanical keyboard in the budget. Almost all the parts have arrived, but the power supply is the last thing scheduled to arrive, so there's not much I can actually do until it gets in on Tuesday. Grrr... I also didn't think I'd say this, but I miss London weather! Boston is super hot right now, and it's making it really hard to focus on getting work done. In addition to having to dodge and/or run over undergrads (especially deer-in-the-headlights freshmen), I have to contend with the hoards of tourists in Harvard Yard.

You're missing some really important info here - namely, what you actually want to work on for a PhD. Mechanical engineering is way too broad for anyone to suggest any schools.

I had 166 Q, 169 V, 3.5 AWA (still no idea what happened there, since I'm generally a strong writer), and I got into all the BME PhD programs I applied to (including Harvard). Once you get to that point, I don't think 1-2 points difference on the GRE are going to make a huge difference. The other parts of your application will matter much more than GRE score at that point. (GPA, letters of recommendation, etc.)

The ease of changing advisors definitely varies by school, but it's not verboten. (I've heard the recommendation to go to a school where there are at least 2 people you could see yourself working with, in case things don't work out with your initial advisor for whatever reason - funding, research area, mentoring style, etc.) With rotations, it's definitely easier to switch advisors since you don't really have one until the second year, but I don't think rotations are common in most engineering fields. In my program, the advisor is "preliminary" for the first year and you have zero obligation to stick with them. However, at another program (CMU) coming in with a masters degree meant that they basically cut off the first year of the program and you were expected to dive in with PhD research right away, making it much harder to change advisors. It's definitely a question to ask when you're doing interviews/visits.