The Eight Coding Interview Patterns That Actually Show Up

There is a long tail of weird coding problems out there, but the head of the distribution is small. If you can recognize and execute eight patterns cleanly, you'll handle 80% of what gets asked in real loops at real companies.

The patterns

1. Two pointers — sorted array problems, palindromes, partitioning.
2. Sliding window — longest/shortest substring, fixed-window stats.
3. Hash map indexing — counting, deduplication, lookups in O(1).
4. Binary search — on sorted data, on answer space, on rotated arrays.
5. Heap / priority queue — top K, scheduling, merging streams.
6. Recursion with memoization — overlapping subproblems, dynamic programming entry point.
7. Graph traversal — BFS for shortest unweighted, DFS for connectivity.
8. Stack — parsing, monotonic stack, balanced expressions.

How to drill them

Pick five to ten problems per pattern, not fifty. After each problem, write a one-sentence note about the pattern, the trick, and the time complexity. Re-read your notes weekly. Pattern recall — not problem recall — is what shows up in interviews.

Skip the rabbit holes

• Bit manipulation puzzles. Almost never asked at companies you'd want to work at.
• Obscure DP variants. Standard knapsack and longest-subsequence is enough.
• Math-heavy problems. If you see Catalan numbers, the company is signaling something off about its hiring bar.

What to spend your last week on

In the final week before a loop, stop learning new patterns. Re-do five medium problems per day across patterns you already know, on a timer, and explain your solution out loud. Speed and verbalization are what break down under pressure — not the pattern itself.