I, a curious person and budding scientist, want to understand more of the world I observe. I make a starting assumption, that it's possible to create a usefully predictive model.

I dream up a model, then I create a set of hypotheses to test the predictions of my model. If my predictions match my observations I say that I have created a useful model. Although you hear people talk about theories being proven, if you peel back what's actually happening it's just creating models. If over the course of all my efforts I never create a useful predictive model I may have disproved my starting assumption that modeling is possible. But we have ample evidence that it is.

Here's an example of a model we know to be untrue (or more correctly incomplete) but which we still use: Newtonian Mechanics. NASA does not need to involve Relativity to send a probe to Mars and have it get there as planned. Because the space-time in which the probe operates is simple enough that Newtonian math is sufficient. But we also have a better model: General Relativity. We have not thrown out Newtonian because its still a useful model provided you know it's limits. Newtonian assumes absolute time and Galilean Invariance. Both models assume the gravitational constant G doesn't change. When/if they find evidence of that we'll need new models.

The idea that the universe, or every other person is not lying to us as you said, is not faith. Just another assumption.

There's no faith here. Just a process. The critical difference between this process and faith is that it MUST be constantly tested and discarded if it does not work.