I haven't interviewed since 2023 and I'm preparing for senior full-stack roles. Are algorithm and LeetCode-style coding exercises still common, or has widespread AI use made them less meaningful? Should I spend time practicing coding puzzles, or focus more on system design, code review, and practical engineering scenarios?
4 Answers
For a senior developer, practical evaluation makes more sense: review flawed code, identify performance and reliability problems, discuss API scaling, and design a maintainable system. A strong senior candidate should be able to recognize when generated code is unsafe or inefficient, not merely produce an algorithm from memory. That said, you may still need basic algorithm practice because some companies haven’t updated their process.
The format hasn’t necessarily become more useful just because AI exists. Some companies still give surprisingly difficult graph or optimization problems for jobs that mostly involve ordinary CRUD work. In those interviews, being comfortable with the common patterns can keep you from being eliminated, even if the exercise isn’t representative of the actual job. Whether a company allows AI varies, so ask about the rules ahead of time if possible.
Yes, it’s still worth doing some preparation because many large companies continue to use algorithmic questions, often as an early pass/fail screen. You probably don’t need to grind hundreds of difficult problems, though. Review the common patterns, practice explaining your thinking, and spend most of your preparation time on system design, behavioral questions, and realistic full-stack scenarios.
LeetCode can still be useful as mental exercise and for refreshing problem-solving skills, but memorizing solution patterns is less valuable than understanding tradeoffs such as time and space complexity. AI can produce code quickly, but you still need enough technical judgment to verify the result and choose the right architecture. I’d use coding puzzles as a smaller part of the plan and prioritize system design, code review, and explaining decisions from past projects.

Exactly. A good exercise could include generated code with subtle bugs and ask what you would change, how it scales, and how you would test it. That reveals much more than memorizing a solution to a puzzle.