The problem with improving seems to be that while I can play a few blitz games, taking time to work on game is more a separate occupation. I occasionally do the lichess practice thingies but without seeing much impact. Anything that can be done that helps but that is on the timescale of a few blitz games?
Whenever there's a position where you don't see the reported advantage, or when a change in the assessment occurs that you didn't expect, that's an opportunity to go back afterwards and just look at those specific points in the game to see what the engine suggests.
If you really want to get good of course you need to be much more methodical in your study. But I started to find chess much more enjoyable once I left behind any long term ambitions.
Which isn't necessarily a bad thing, it just changes the game. If we embrace this rule, then I humbly suggest that a large change in evaluation (due to a blunder) be accompanied by an "Oooooooh!" sound effect; that's what semi-restrained human observers might do.