I've done the leetcode thing. As a hiring manager, I don't find much value in homogenization of skillsets that comes from dedicating time to memorizing rote answers to problems. Some amount of that skill on a team can be beneficial, but it's less effective when everyone brings that skillset to the table.

The issue with leetcode is a lot like the issue of GMAT/GRE/SAT/ACT. It can be used to remove candidates from the pool, but it is a really poor indicator of general success overall. All it does is make the candidate pool a bit more manageable at the expense of people who have other skillsets.

Skillsets that are selected against by leetcode are tip-of-the-spear adopters and other innovators (broad thinkers), communication skillsets, and often UX understanding/projection.

As an employer, I have used Hackerrank for a specific skillset and role we needed (deep SQL experience), but that is the only type of testing we used for selection processing.