Why LeetCode doesn't fit most roles
LeetCode is great for one thing: algorithm interviews for software engineers. But a product manager gets asked why a metric dropped. A data analyst writes SQL against messy tables. A finance associate builds a model. None of that is a linked list.
Crossway Arena
Arena on Crossway has real work problems for every field: metric drops, A/B test reads, pricing and market sizing for product, SQL and statistics for data, P&L and DCF for finance, pipeline math for sales, and coding and system design for engineering. Problems go from Easy to Hard, are graded instantly, and each solve shows on your profile and lifts your job matches.
For SQL and data
StrataScratch and DataLemur both focus on SQL and data interview questions, many modeled on what companies ask. Pair them with practice on messy, real looking data, since that's what the job actually is.
For product
Product interviews test product sense, metrics, execution and estimation. Practice out loud: pick an app, define its north star metric, then explain what you'd do if it fell 10% overnight. Arena's product problems give you the numbers and check your answer.
For consulting and finance
Case interviews reward structure and fast math. Practice market sizing and profitability cases with a timer, and build a few small models from scratch: a P&L, a break even and a DCF. Speed comes from reps, not reading.