Review & Improvement

Quality matters more than quantity. Solving 150–200 well-chosen problems with thorough review is more effective than grinding 500 problems without understanding patterns. Focus on covering all major patterns (hash map, two pointers, sliding window, BFS/DFS, binary search, DP, backtracking, heap, sort/greedy) with 10–15 problems each. If you can solve a new Medium problem in 20 minutes consistently, you are ready.

No. Memorize patterns, not solutions. If you memorize "Two Sum uses a hash map," you can solve any variation (Three Sum, Four Sum, subarray sum). If you memorize the exact Two Sum code, you will fail when the problem is slightly different. Focus on understanding why a pattern works, not what the code looks like. The code should flow naturally from understanding the pattern.

Hard problems usually combine 2–3 patterns. The time sink is typically in identifying the right approach, not coding it. Practice: (1) Read the problem and write down 3 possible approaches in 2 minutes. (2) For each approach, estimate the time complexity. (3) Pick the one that meets constraints. Also, make sure you can code Medium problems in 15 minutes or less — that gives you 30+ minutes for the Hard problem.

Use these techniques: (1) Rubber duck debugging: explain your approach out loud as if teaching someone. This reveals gaps in your understanding. (2) Record yourself: use screen recording while solving problems. Review the recording to see where you waste time. (3) Use platforms: Pramp, Interviewing.io, and LeetCode contests simulate real interview pressure. (4) Write solutions from scratch: after studying a solution, close it and rewrite from memory.

This is completely normal and expected. The forgetting curve is steep: you will forget 80% of what you learn within a week without review. That is why spaced repetition is essential. When you forget a pattern: (1) Do not panic — re-learning is faster than initial learning. (2) Re-read the pattern explanation (not just the code). (3) Solve one problem using that pattern from scratch. (4) Schedule another re-solve in 2 days. After 3–4 re-solves over 2 weeks, the pattern will stick permanently.

Use whatever language you are most fluent in. Python is popular because of its concise syntax and built-in data structures (Counter, defaultdict, heapq, deque). Java and C++ work well too but require more boilerplate. The key is being able to translate your approach into code quickly without fighting the language. Do not switch languages just for interviews — the overhead of unfamiliarity will slow you down more than any language advantage.

In a real interview, yes — mention the brute force approach and its complexity, even if you do not code it. This shows you understand the problem and can analyze trade-offs. In a timed exam, code the brute force only if: (1) you cannot think of the optimal approach after 5 minutes, or (2) you need a working solution for partial credit. If you immediately see the optimal approach, go straight for it and mention "the brute force would be O(n^2) but we can do O(n) with a hash map."