How to (try to) turn an MS into (a shot at) a PhD (if you must).
Warning: this is advice I have not tried, and I would not expect it to have worked for me. It is, however, my best guess about how to create an unofficial path where one otherwise might not exist.
Shriram Krishnamurthi’s FAQ on getting a CS PhD in the US is the best general advice I know. What follows is about the narrower problem of trying to use a US MS program as preparation for a PhD.
That usually means sustained work on an uncertain problem, some artifact that makes the work visible, and a detailed letter from a researcher who watched you do it. A thesis, manuscript, system, or careful reproducibility result may supply the artifact.
This advice is for people who have not already been actively engaged in a research lab, including people whose undergraduate program did not give them that opportunity.
The difficulty
At least in the US, an MS is not a small PhD. Standalone MS programs are course-based professional degrees.
Unless stated otherwise, MS programs are ‘terminal’ and not designed to turn their students into PhD applicants. I would not expect the admissions staff, or even most faculty, to know how an MS student can make that happen. This is an uncommon path, and you may have to create it for yourself.
If this is nonetheless your quest, then make this the minimum test of any MS program you consider: can students do independent study and write a thesis as part of the degree?
I would apply to MS programs almost as though you were applying to work with particular PhD advisors, then try to create a shadow apprenticeship before you arrive. Choose programs around particular researchers, not simply departments or course offerings.
Read enough of their recent work to understand what their groups are actually doing, look at their students and research software, and begin writing to people before you apply. You may have better luck by writing to graduate students than to professors. Graduate students are often closer to unfinished work and may have a small subproblem on which help would actually be useful.
Everything I just said, except the graduate-student part, is what you would have been doing to find a PhD position. Because you’re doing essentially that.
“Free labor” alone isn’t sufficiently attractive: the scarce resource is time and supervision. You need to offer free labor whose expected supervisory cost is low—reproducing a result, getting a research system running, handling an open issue, running experiments, improving tooling, etc. Be reliably useful enough that involving you becomes easier than not involving you.
You may have to send a lot of mail. Make each message specific enough to show that you have read the work, follow up once, and expect most attempts to go nowhere. If someone gives you an opening, do a killer job, and ask for more.
The message should be cheap to evaluate. In a few sentences: say what work of theirs you read or tried, identify the particular thing you could do, and ask whether that contribution would be useful. Attach evidence only when it helps them judge the offer: a small repository, reproduction note, bug report, or short technical write-up is better than even the most enthusiastic fan letter.
The goal need not be a publication; it should be serious enough to produce a thesis, manuscript, system, or other clear artifact, and to let an advisor write in detail about you.
You want to be able to earn a letter from this person. You want to demonstrate enough capability and talent that they would recommend their personal researcher friend at another school invest 300,000 and a half a decade in you.
Remember that if the MS is only a year and you apply to PhD programs during your first fall, you would have only a few months in the program to build those relationships and results. Even if you’ve had a 6-9-month collaboration before starting the MS, it may still be worth choosing a two-year program or allowing an extra application cycle rather than rushing the very part of the master’s you need most.
The attempt is to manufacture the missing apprenticeship: find the people doing work you care about, make it easy for one of them to try working with you, learn fast, become genuinely useful, and turn that usefulness into visible research judgment.
This advice asks you to behave like a beginning graduate student before you have received the apprenticeship that normally teaches someone how to be one. This is an unfairly high bar, and it is why I would not have expected this strategy to work for me. Nonetheless, I think this is the most plausible way to turn an MS into a PhD apprenticeship.
If you can choose a route that already provides research mentorship, choose it. If an MS is the route you must take, coursework is the official program. The research apprenticeship is the unofficial program you are trying to build inside it.