I'm probably going to be one of many who will say that you should pick a pain you're already having and try to solve it. It's the only way you'll know exactly what to build and how to build it.

The more you try this the easier it will become, and after a few failed attempts you'll know how to get from the "new idea" to a finished product. Yes failed attemps, it'll probably take a while until you learn which things are worth building and which aren't.

Don't be afraid to be very specific. Pick a niche problem, don't try to build a solution that fits everyone. Just go after one specific use case and do it right.

You can try solving someone else's problem, or a problem you see in the company you work for, or a problem your customers are having. But be sure to understand the pain you're trying to solve, feel the pain, you need to know why are you building this thing. Don't just do it because it is cool, do it for a reason.