I would like to know as to how or at what rate does the academic life or a good graduate student progress in a theoretical physics grad school if he/she is aiming to specialize in topics like say the Maldacena conjecture (or "AdS/CFT" as it is often said), or issues of integrability emerging from it and such related questions. I have in my mind "role models" of certain brilliant students who have recently completed their PhDs in such stuff like Jacob Bourjaily, Tudor Dimofte, Silviu Pufu etc.

Assume that a grad student starts learning QFT and related mathematics from the first day of grad school (which I guess is already too late!)
I guess some of these recent very successful PhDs were way ahead of such scenarios!

Then, how much and how far into QFT should someone know before he/she can get into cutting edge of Maldacena's conjecture?

To split up the question,

How long should it take to learn enough of QFT (what is enough?) ?

At what stage and after knowing how much of QFT should one be able to start learning String Theory?

How long and how much of String Theory should one know before being able to get through the literature in the topics mentioned above?

Can one start reading the papers (or even working?) in these topics along with learning QFT?

Can one start learning String Theory along with learning QFT or do they have to come in a strict order?

I guess the canonical resource to learn about Maldacena's conjecture is the famous "MAGOO" review or may be the Hoker and Freedman's review.
Any other suggestions?
The "problem" I see is that standard QFT courses don't teach about N=4 SYM or about N=8 Supergravity and absolute fluency with such stuff seems to be utterly necessary for getting into this field.

For all of these above questions I would love to know of the characteristic time-line in terms of months into graduate school when each of these milestones should be covered.

I guess this will help know what is the right pace to work at - which I guess was the speed in which some of these people mentioned earlier worked at.

This post has been migrated from (A51.SE)