That said, this reward style finds significant problems in the literature. Expected rewards act like a chemical fix (quite literally, in terms of how they work in the brain), that require larger and larger doses over time, and eventually crowd out the will to do the behavior on your own. I'd be interested to see how using something like this would work long term. I know its rather simple in its current form, but so are most of the examples researchers use in the laboratory.
Regardless of the science, great job on finishing and launching your first Rails app. My only technical note: your images are a bit pixely on my Retina display. Other than that it looks great!