Coding challenges have been the standard for evaluating candidates, but are they actually a good predictor of on-the-job performance? Yes, they test problem-solving skills and algorithm knowledge, but let’s face it, most of us don’t write binary search algorithms daily. The pressure of whiteboard interviews? Not helpful.

Some companies are starting to shift focus to more relevant scenarios—like working on real projects, pair programming, or system design discussions. Makes sense. Would much rather see how someone approaches a problem with realistic constraints than have them memorize LeetCode patterns. But then again, how do you balance that with gauging fundamental coding skills? It’s a tricky thing to figure out.