Mock Interview — Onsite / Hard

Onsite Strategy

Onsite problems are less about pattern recognition and more about combining multiple patterns, handling tricky edge cases, and optimizing beyond the obvious solution. Expect 35-45 minutes per problem.

• Spend 5-7 minutes clarifying — on hard problems, ambiguity is often intentional. Ask about constraints, expected output format, and edge cases thoroughly.
• Start with a feasible approach — even if it's O(n²). Hard problems often require getting the brute force right first, then optimizing.
• State the lower bound — say 'This is O(n log n) because we must sort, so O(n) isn't possible.' or 'We must examine every element, so Omega(n) is the lower bound.'
• Optimize incrementally — O(n²) → O(n log n) → O(n) → O(log n). Show your thought process stepping through each improvement.
• If stuck, simplify — solve a restricted version (e.g., assume no duplicates, assume sorted input, assume tree is balanced) then generalize.

Graph & Tree (Hard)

These test your ability to combine traversal with state tracking, parent pointers, or advanced tree properties.

DP & Optimization (Hard)

Hard DP problems often require state compression, non-obvious state definitions, or combining DP with another data structure.

Advanced Data Structures (Hard)

These test your ability to design and combine data structures under constraints.

Key Points

• Hard problems test process over perfection. A working O(n²) solution is better than a broken O(n) one.
• Use the full 45 minutes. Don't rush. Pause, think, then speak.
• Interval DP (burst balloons, matrix multiplication) is a distinct sub-pattern — practice it separately.
• Advanced data structure problems (LFU, All O(1)) test whether you can combine multiple primitives correctly.
• After the interview, write down what blocked you. Was it the recurrence? The data structure choice? The edge case? Target those weak points.