Like my past studying for SAT or LSAT, it was as much about preparing myself for the test itself vs. trying to gain some specific knowledge. Time management, making sure to test, and knowing that there are almost always naive vs. optimized solutions for these types of problems and learning how to quickly identify the naive solution (i.e prove I can solve) and then figuring out the optimization.
In any case, I enjoyed the process and didn't really practice more than 10-15 hours. I literally just kicked off the rust and got used to managing time. If you get too deep in to "I've optimized systems to process millions of transactions/second, and some 25 year old is sweating me on hashtables" you're going to have a bad time.
The ironic part of the interviews at this level of experience was that the coding portions ended up feeling tough, but I was able to talk work through them. For one company, one of the two system design round was with someone who couldn't have been more than 3 years out of college (with PhD though) and had only been at that big company. That was my hardest interview because the person I was talking to had zero to little actual experience there. They knew the problem and expected answer, but didn't know the space. It was VERY HARD to talk about things that they very obviously didn't understand in terms that they did understand without sounding condescending.
I still don't know if I'm reading amen or a man.
But I hear these stories over and over - the variance is so wildly high. The people who spent the same or more hours studying and didn’t get an offer tend to not talk about it though. So, sampling bias issue.
Have not done a leetcode-style interview yet, and might never do, and might not pass if I do do, but just as a scratching post for your coding claws I think it's awesome.
I guess where I'm going with this is, maybe change your thinking about leetcode? Don't think of it as a necessary evil in order to land a high paying job that you dread doing each night; look at it as a fun little hobby, with the nice side-effect that you're keeping your data structure knowledge, algorithms and general problem solving skills sharp.