Two Sum
Return indices of two numbers adding to a target. The one-pass hash map is the expected O(n).
Worked Amazon set →LRU Cache
Design get/put in O(1). Hash map + doubly linked list is the canonical answer.
Number of Islands
Count connected land cells in a grid. DFS/BFS flood fill or Union-Find.
Merge K Sorted Lists
Min-heap of list heads, O(n log k). A classic Amazon hard.
Trapping Rain Water
Two-pointer sweep tracking left/right max. Often the second-round closer.
“Tell me about a time you disagreed with your manager.”
Tests Have Backbone; Disagree & Commit. Use a STAR story with data and a respectful resolution.
Amazon behavioral guide →“Tell me about a time you failed.”
Tests Ownership + learning. Own it, show the fix and what changed afterward.
Design a Notification System
Multi-channel (push/email/SMS), fan-out, retries, rate limiting, and dedup.
Worked solution →Median of Two Sorted Arrays
O(log(m+n)) binary-search partition. A signature Google hard.
Worked Google set →Word Ladder
Shortest transformation sequence — BFS over a word graph.
Course Schedule
Can all courses be finished? Topological sort / cycle detection in a DAG.
“How do you handle ambiguity?”
Google’s “Googleyness” signal — show structured progress without full information.
Design a Web Crawler
URL frontier, politeness, dedup, distributed fetch — classic Google design.
Worked solution →Valid Palindrome
Two pointers from both ends. Often paired with a second harder problem.
Worked Meta set →Subarray Sum Equals K
Prefix sums + hash map of counts. A Meta favorite.
Merge Intervals
Sort by start, merge overlaps. Tests clean interval handling.
“A time you influenced without authority.”
Meta values cross-team impact. Show how you aligned people without owning them.
Design a News Feed
Fan-out on write vs read, ranking, and the celebrity problem.
Worked solution →Design Twitter
Timeline generation, fan-out, and read/write trade-offs at scale.
Worked solution →Reverse a Linked List
Iterative pointer re-linking, O(n)/O(1). A Microsoft warm-up staple.
Spiral Matrix
Boundary-shrinking traversal. Tests careful index bookkeeping.
Clone Graph
DFS/BFS with a visited map from original to copy.
“Tell me about a conflict with a teammate.”
Show empathy, a concrete resolution, and what you learned about collaboration.
Min Stack
Support push/pop/top and min in O(1) — keep an auxiliary min stack.
LRU Cache
O(1) get/put. Apple often digs into the data-structure trade-offs.
“A time you pushed for a higher quality bar.”
Apple prizes craftsmanship — show where you refused to ship something half-done.
Design Video Streaming (Netflix/YouTube)
Upload/transcoding pipeline, CDN egress, and adaptive bitrate.
Worked solution →Design a Distributed Cache
Consistent hashing, eviction, replication, and hot-key handling.
“A high-judgment call you made independently.”
Netflix’s “freedom & responsibility” bar — show senior, independent judgment.
Optimize Matrix Multiplication
Cache blocking and memory-access patterns — Nvidia probes performance thinking.
Worked Nvidia set →Producer–Consumer (Threads)
Bounded buffer with mutex + condition variables. Tests real concurrency.
Design a URL Shortener
Key generation (base62), redirect path, and read-heavy caching.
Worked solution →Design a Rate Limiter
Token bucket, distributed counters in Redis, and edge cases at scale.
Worked solution →Design a Chat System
WebSocket delivery, presence, ordering, and offline message sync.
Worked solution →No questions match those filters. Try clearing the search or picking “All.”
FAQ
What types of questions do FAANG interviews ask?
Coding (DS&A, usually 2 rounds), system design (open-ended, mid/senior), and behavioral (your experience vs the company’s values — Amazon’s Leadership Principles being the most famous). Filter this bank by both company and type.
Are these the exact questions companies ask?
They’re representative, widely-reported questions in each company’s style. Companies rotate exact questions — practice the formats and patterns, not memorized answers.
How should I practice?
Pick a company, filter to one type, and do timed attempts out loud: restate the problem, talk through the approach before coding, state the complexity. For behavioral, prepare adaptable STAR stories.
Is it free?
Yes — free, no signup, filter on any device. Published by CoPilot Interview.
Practice is good. Game-day help is better.
When the real round starts, CoPilot Interview is a desktop AI assistant that gives real-time, structured help — approach, code, Big-O, and STAR scaffolding — during the actual interview, with a permanent free tier.
See how it works