I use usearch for the engine https://github.com/unum-cloud/usearch/.
It's fast. The vectors are stored in DB and added to usearch at startup for retreival.
There is also FAISS.
The way I think is the most crucial thing in my case.
I tend to focus more on the positive (I trained myself) and being healthy (food, sport, sleep, walks...). With good energy it's way easier.
From what you wrote, I'm not sure, but it seems that the "talking to people" and the "problem definition" steps could be improved to better identify the SPECIFIC problem you are solving, the causes behind the problem, the users motivation to solve it... By asking questions like: Is it a problem people are trying to solve? How many? At which frequency?
Because solving an important and specific problem will make it easier to target potential users and for them to be naturally interested.