This may be won't be a answer, you expect. But, nevertheless.
Without wanting to turn some one down, I would try to get an answer for these questions firstly:
- why are you "deciding" for your girlfriend, what (c/w)ould be good for her. Isn't it her invested time, degree and latter responsibilities@paperwork?
- she studied that, so she should have some ideas what is possible on that topic. Why are you trying to find something (..of course, one love, one heart)
- I don't have a CS background and I learned to python by my self. Or, better said, how to copy+paste. It was a pile of work an a lot of time went into understanding, debugging, refactoring.
Could your girlfriend do that? Is she willing to learn and eager to do that?
I don't know you guys, but I know quite a lot of PhDs and from what I can tell, if one is a physicist, it's really difficult to talk about chemistry with him/her as both don't speak the same language and neither even understand whats been talked about.
For a PhD in medicine it could be the same with programming.
So, this questions need to be answered upfront.
It's important to understand, that by "I can help you, I know good thing..may be you can do so.." we mostly impose hard work and time and pressure on the others, especially when the studies done are completely different in topic.
The next thing I would do is to start a research on the latest findings in that topic and just look at the methodology used for that papers.
If at that point both of you think "what the hell", then may be programming is not for her.
You should've experienced how it feels to understand code written by others - in principle, you're capable to speak & understand that language. But stil, all the logic. ...
So if she is struggling with other's methods of concluding smth on her topic, than it w/could be difficult for her to understand all the programing techniques and the logic behind e.g. A2C nets in reinforced learning.
For me it's difficult as hell to (!)understand(!) what's happening, because I don't have a mathematical, statistical background. So will your girlfriend..?
Found the latest findings on her field you can derive Information out of the papers what is used and what skills are used for that.
Usually, the papers are written under time pressure, so I think you can derive state-of-the-art methodology (where the writers don't have time to learn a new language and produce good code, but instead use what they have learned while studying). This will be the "basic skills set".
Of course, doing so, one need to filter A LOT A LOT..... Feasibility is also a good term in here.
By doing so, you'll not just collect ideas, skill necessaties, but also collect knowledge and she should be able to assess that with her knowledge.
And then, a year or so has passed, you'll know where she wants to head to :)
It's little bit difficult to find ideas and how-tos. They will emerge while one is dealing with something.
But, a programmers approach that is also used by me in daily life:
- Code, run, debug, repeat
- delete as much code as possible
Can be applied to everything. So, I would start with AI and get all the latest news on usage of AI in cardio. Having that, get the return of the function "how my knowledge fits into", pipe that to function "what is needed" - and delete all the code as it's none sense.
I think, you understand what I'm trying to say :)
There will be a lot of work. For both of you. But both gain knowledge.
:) I like