[1123581321]

I think your best bet is to keep pursuing the "building things" strategy, but when you get stuck on something, look for exercises or deep reading in that particular area. For example, awhile ago I got stuck on geocoding and geosearch and so went in search of tutorials and resources about that particular thing. I then took that information and went back to building my app.

I think it's a much better use of your time than doing general intermediate exercises without specifying a particular area in which to learn. At the intermediate level, your options expand so much that no procedural resource can both cover what troubles you specifically and also challenge you consistently.