I'm in a similar position. I built a thing that I'm turning into a startup, but I'm coming out of the military and don't know anyone in a position to join me on a startup, let alone interested.
The first strategy I'm working with is going through the short list of people I trust and trying to get them interested enough in the vision to think about joining me. I've had some success with this approach, but results are insufficient. I assume that what I really need is commitment to the vision and that most of the skills can be acquired. I could be wrong about that.
The secondary strategy I'm working with is reading as much as I can possibly find on the various subjects related to my startup and its domain, and keeping track of the people who stand out in some way. This is limited by the fact that people have to be publishing or I won't find out about them.
The tertiary strategy I'm going to try is pursuing all of the vaguely warm connections I might have to different groups, like veterans groups, alumni groups, entrepreneur groups, etc. At least having that factor in common will help.
The backup strategy is to just do a normal search of the marketplace for anyone I might like to work with and hustle my way into meetings or working together. Maybe something will develop, but this strategy isn't significantly different from the default...
...which is to just be a single founder. Others have made it work. It's not ideal, but it shouldn't make things impossible either.
Anywho, with respect to your specific situation, it seems like it would make sense to move to NYC. It's got a good startup environment and already has a focus on finance. You could try to just cold email a few dozen well connected people in NYC and say that you're going to be in town and you want to talk about startups for 15 minutes. If nobody bites, then don't go. If you get a bite, buy the ticket and then email more people and say you're going to be in town to met with so-and-so and while you're there you'd like to chat about startups for a minute.
I haven't gotten a chance to use that approach myself yet, but apparently it works.