What to do after being rejected from every PhD program?

[naikin]

I applied for PhD in AI and got rejected from every program (Berkeley, Stanford, CMU, UWash, MIT, Georgia Tech). I am just finishing my EECS Masters in the UK. People told me that my Statement of Purpose was very strong. My recommenders are not famous in the US - one is working in Robotics, another in AI and another in CS. I have not seen the recommendations, but the robotics one must have been good since it got me to the interview stage for Google AI Residency. I have seen a previous recommendation from the CS professor and it was also very good. At the time of application I had a robotics paper published to ICRA (second author) and was working on another AI paper.

I want to try again next year, but I am not sure what to do in order to improve my record. All PhD programs refuse to give me any feedback. I have a feeling that doing research under the supervision of a famous AI researcher will be probably the best option. However, getting a famous US researcher to work with a random person (not even a student anymore) from another country seems extremely unlikely. I got to the last stage of the Google AI Residency but got rejected in the end (the recruiter hinted that other applicants had stronger background). Also got rejected for a research position at DeepMind after several interviews (feedback was that my AI knowledge was very good but coding skills were on the boundary).

Now that I am graduating, I have to find a job to support myself. I just do not know what to do in order to stay involved in AI research. More specifically, I want to work in Reinforcement Learning applied to robotics. Any suggestions are welcome. Thanks!

[naikin]

Just some clarification on my original question. I am particularly interested in advice on what to do the after graduation and before the applications for next year are due. Specifically, I find it difficult to get a research related AI job since the industry is still mainly dominated by people with a PhD (although I have not really applied to that many places). I am also interested in possible suggestions to stay in academia even after graduation, since this will allow me to continue doing research. Ideally, it would be great if I manage to get a paid AI research internship in some good US university, but I have no idea if this is possible since I will no longer be a student and I will also need a VISA. Thanks!

[petercooper]

Why not do your PhD in the UK? There are lots of good AI programs here - some world class (Imperial, Edinburgh, UCL), some not quite so.

[Broken_Hippo → petercooper]

To expand on this, they might try other countries as well as the PhD programs might be in English even though the country isn't a predominantly English speaking country. For example, I know that many of the Masters and PhD programs here at NTNU in Trondheim, Norway are English instruction. Other universities (here and elsewhere) might have similar programs.

I have no clue about quality of programs (world class vs not) but figured I'd suggest the option.

[naikin → petercooper]

Thanks, for the suggestions. Next year I will definitely try UK PhD programs. Nevertheless, I am still very biased towards the US for several reasons: 1. It has been a dream for me for a long time and I know that eventually I want to go study there. 2. The professors who are doing work that best matches my interests (RL applied to robotics) are mostly located in the US. The best match I would say are people such as Sergey Levine and Pieter Abbeel. Others are Anca Dragan, Dieter Fox, Emo Todorov, Drew Bagnell, Abhinav Gupta, etc. 3. The good universities in Europe which have researchers who are actively working in RL for robotics are not that many. Oxford is maybe the closest match. While the other universities mentioned also have very good people working in AI, their focus is usually either too theoretical for my interests or oriented towards pure ML and DL.

I know this might sound too picky and fastidious, but bottom line is that I will be spending approx 5 years in this program, going very deep and working very hard. It is just going to be a failed PhD if I take a position just for the sake of doing a PhD and end up doing research that I am not really interested in.

[nextos → naikin]

Oxbridge has admission routes where you get in touch with professors directly, and if interested they will offer you a position using their own funding. Thus, you don't need to wait till next year, nor applying to a specific programme. This happens anytime in the year.

[nextos → petercooper]

Cambridge and Oxford have excellent AI people too (e.g. Ghahramani now on leave from Cambridge to Uber, or Frank Wood in Oxford).

Besides, some Scandinavian universities have excellent professors, they will pay you a great salary (not a stipend!) and you will graduate quickly. I'd rather look for an excellent advisor than for a great university if I had to compromise a bit.