I'd recommend easing yourself into an existing technical community organized around something you use or find interesting. (This will probably be software/hardware, but it could honestly be any field of knowledge you find interesting--there are inevitably anthropologists, biologists, painters, photographers, etc. out in the world who wish they had some tool they don't know how to build.)
To start, just explore, learn, and help others. (If you find this boring, it's a good sign to pick another community.) Sooner or later, you'll bump into real bugs and missing features/tools, and you'll have an excuse to start learning how to implement them. If you feel like you are understanding very well (perhaps you're able to regularly help people from memory) and no good opportunities have presented themselves, you could also start asking around. I would focus on people who are visibly working on their own projects; ask for any standalone projects on their todo lists that they don't have time/energy for.
---
As an aside, the right project probably depends on what motivates you. Not what motivated you to post this--but what motivates you to keep waking up on this side of the dirt each day. I.e.:
- If you are motivated by learning more than building, you may be able to get away with picking something at random.
- If you are more motivated by building things (a good test--does leaving things unfinished weigh heavily on you?), you'll probably do best if you can find something that will scratch an itch of your own. When you're scratching your own itch, it's easier to avoid analysis-paralysis over the needs of users that you don't have yet. Unfortunately, there's a chicken/egg problem here. It's easier to find itches when you're already working on some kind of project (of your own, or with others). I regularly bump into little things that could make a good project.
- If you are indeed more motivated by helping other people, it's good to start preparing now for the possibility that you are in for a rough ride. There are a lot of missing little tools out in the world that would be useful for other developers or the general public--but there's also a vast sea of interesting little tools that've already been built that have somewhere between 0 and 100 users. You will not only have to solve the problem, but to help others you'll probably have to promote your own work. (Pay attention to whether this sounds boring, exciting, or horrifying. You can overcome boring/horrifying if you are really committed, but it may be a grind if you aren't...)