Skip to content

Hard Roadmap

The differentiator for Senior/Staff roles.

Introduction

"Hard" LeetCode questions are rarely asked for entry-level or junior roles, but they are increasingly common for Senior, Staff, or highly competitive Quantitative Finance roles.

Hard questions are usually one of two things: 1. Two completely different Medium patterns chained together. 2. A single extremely obscure algorithm (like A*, Kosaraju's, or complex Bitmask DP).

The Goals of Hard Questions

  1. Managing Complexity: Writing 60-100 lines of bug-free code without losing track of your state.
  2. Advanced Data Structures: Using Segment Trees, Tries, or Disjoint Sets (Union Find).
  3. Advanced Optimization: Reducing \(O(N^2)\) to \(O(N \log N)\) or \(O(N)\) using Monotonic Stacks or Deques.

The Hard Patterns to Master

Only study these if you are consistently crushing Mediums.

1. Monotonic Stack

Maintaining a strictly increasing or decreasing stack to find the "Next Greater Element" in \(O(N)\) time. - Classic Problem: Trapping Rain Water (LeetCode 42) - Classic Problem: Largest Rectangle in Histogram (LeetCode 84)

2. Union Find (Disjoint Set)

Finding connected components and cycles in undirected graphs efficiently. - Classic Problem: Redundant Connection (LeetCode 684) - Classic Problem: Number of Connected Components in an Undirected Graph (LeetCode 323)

3. Tries (Prefix Trees)

Storing strings character-by-character for lightning-fast prefix lookups. - Classic Problem: Implement Trie (LeetCode 208) - Classic Problem: Word Search II (LeetCode 212) - Combines Trie with Matrix DFS!

4. Advanced Graph Algorithms (Dijkstra / Topological Sort)

Finding shortest paths in weighted graphs, or ordering graphs with dependencies. - Classic Problem: Network Delay Time (LeetCode 743) - Dijkstra's - Classic Problem: Alien Dictionary (LeetCode 269) - Topological Sort

5. 2D Dynamic Programming

State depends on two changing variables (usually represented by a 2D matrix). - Classic Problem: Longest Common Subsequence (LeetCode 1143) - Classic Problem: Edit Distance (LeetCode 72)

6. Median / Sliding Window Heaps

Using TWO heaps (a Max Heap and a Min Heap) simultaneously. - Classic Problem: Find Median from Data Stream (LeetCode 295) - Classic Problem: Sliding Window Maximum (LeetCode 239)

The Reality of Hard Questions

If you get a Hard question in an interview, the interviewer is often testing your problem-solving resilience more than your memorization.

They expect you to struggle. They expect to give you hints. If you communicate clearly, implement a solid brute-force, and work collaboratively with them to optimize it using a complex pattern, you will pass.

Key Takeaways

  • Hards often combine two Medium patterns.
  • Master Monotonic Stacks and Union Find.
  • Do not let a Hard question panic you; communicate and build up from the brute force.