To really understand what's going on you, learn your tech stack at least 1 layer of abstraction lower than the layer you typically work in. The biggest wins ar

To really understand what's going on you, learn your tech stack at least 1 layer of abstraction lower than the layer you typically work in. The biggest wins are where you make a good decision that avoids an entire class of bugs or an avoid having to write an entire layer of software. These are also the hardest wins to measure. Pick something you are genuinely interested in and learn it very, very deeply. Don't jump around constantly or you won't have insight. One thing I wish school taught was code maintenance. Most of the code you write has a shelf life of a few weeks as you fulfill your assignments. Bad decisions only come to haunt you after months. You need to watch for new technologies that matter, but 90% of what is new will not matter. One technique I've seen work is to build a list of people in the industry you respect and watch what they're interested in.