LinkedIn Coding Interview Questions
32 LinkedIn coding interview problems with full optimal solutions — 4 easy, 21 medium, 7 hard. Every problem ships with multiple approaches (brute-force first, then the optimal), complexity tables for each, company-specific tips on what an LinkedIn interviewer values, and a FAQ section.
- #20easyfoundational
20. Valid Parentheses
Determine whether a string of brackets is correctly matched using a stack — LinkedIn uses this as a 5-minute coding warm-up, often paired with a JSON or template-syntax validation scenario to tie the abstract stack to the profile-rendering pipeline that parses nested markup in member summaries.
- #56mediumfoundational
56. Merge Intervals
Merge all overlapping intervals in a list into non-overlapping ranges — LinkedIn uses this pattern to consolidate overlapping employment date ranges in member profiles and to compact calendar availability windows, making it one of the most practically grounded problems in their interview bank.
- #91mediumfoundational
91. Decode Ways
Count the number of ways to decode a digit string where each letter maps to 1–26 — LinkedIn asks this DP problem to test whether you can enumerate valid interpretation paths under constraints, the same combinatorial reasoning used in their NLP pipeline for ambiguous entity recognition in profile text.
- #128hardfoundational
128. Longest Consecutive Sequence
Find the length of the longest run of consecutive integers in an unsorted array in O(n) time — LinkedIn uses this problem to check whether you can design set-based O(n) algorithms, the same reasoning they apply to detecting unbroken skill-progression sequences in member credential timelines.
- #133mediumfoundational
133. Clone Graph
Deep-copy a connected undirected graph using BFS and a hash map — a near-literal model of how LinkedIn clones member connection graphs across data centers, making it one of the most contextually on-brand questions in their rotation.
- #199mediumfoundational
199. Binary Tree Right Side View
Return the values visible when looking at a binary tree from the right side — LinkedIn applies this BFS-by-level pattern to render hierarchical org-chart data, surfacing only the rightmost node at each reporting tier, a frequent warm-up before harder tree design questions.
- #200mediumfoundational
200. Number of Islands
Count connected landmasses in a grid using BFS or DFS — LinkedIn uses this to probe how you think about connected-component discovery, the same graph reasoning behind mapping professional networks and finding isolated clusters in the member graph.
- #207mediumfoundational
207. Course Schedule
Determine if a set of courses with prerequisites can all be finished — this is cycle detection on a directed graph, the exact algorithm LinkedIn runs to validate skill-path dependency chains and surface circular credential requirements in their learning platform.
- #238mediumfoundational
238. Product of Array Except Self
Compute the product of every array element except the current one — without division — using prefix and suffix passes. LinkedIn asks this to test whether you can reason about O(n) algorithms under artificial constraints, mirroring how their data pipeline computes per-member aggregates that exclude the member's own data.
- #253mediumfoundational
253. Meeting Rooms II
Find the minimum number of meeting rooms needed to schedule all intervals without overlap — LinkedIn applies this algorithm to allocate recruiter calendar slots and to detect overlapping employment periods in member profiles, one of their most-cited interval problems.
- #295hardfoundational
295. Find Median from Data Stream
Design a data structure that computes the median of an ever-growing stream using two heaps — LinkedIn runs this pattern to rank feed items by engagement score in real time, where you need a running median without resorting the entire dataset on every new signal.
- #347mediumfoundational
347. Top K Frequent Elements
Return the k most frequent elements from an array — this is the algorithm behind LinkedIn's 'People You May Know' ranking and trending job suggestion surfaces, where you need the top-K signals from a frequency count without sorting the entire dataset.
- #1easyfoundational
1. Two Sum
Two Sum is LinkedIn's canonical 10-minute warm-up: given an integer array and a target, return the indices of the two numbers that add up to target. The interviewer is grading your willingness to narrate brute-force first, then move to the hash-map optimization.
4 free resourcesSolve → - #47mediumfrequently asked
47. Permutations II
Given a collection of integers (possibly with duplicates), return all unique permutations. LinkedIn asks this on the backtracking round — they want the sort-then-skip-duplicates pattern, not generate-then-dedupe.
4 free resourcesSolve → - #68hardcompany favorite
68. Text Justification
Given an array of words and a max line width, fully justify each line (with the last line left-aligned). LinkedIn asks this for the hard slot because it's the canonical 'simulate carefully' problem — no clever algorithm, just precise spacing distribution.
4 free resourcesSolve → - #76hardfrequently asked
76. Minimum Window Substring
Given strings s and t, return the smallest substring of s that contains every character of t (with multiplicity). LinkedIn asks this on senior loops because the canonical sliding-window template is small to write but easy to break on duplicates.
4 free resourcesSolve → - #126hardfrequently asked
126. Word Ladder II
Given begin and end words and a dictionary, return ALL shortest transformation sequences. LinkedIn asks this for the hardest BFS slot — the trick is to do BFS for distances first, then DFS to reconstruct paths only following shorter-distance neighbors.
4 free resourcesSolve → - #127hardfrequently asked
127. Word Ladder
Given begin and end words and a dictionary, return the length of the shortest transformation sequence where each step changes one letter and produces a dictionary word. LinkedIn asks this on the BFS round — they want the wildcard-bucket preprocessing trick for O(N * L^2).
4 free resourcesSolve → - #150mediumcompany favorite
150. Evaluate Reverse Polish Notation
Given an array of tokens in Reverse Polish (postfix) notation, evaluate and return the result. LinkedIn asks this because it's the textbook stack problem — they want clean push-on-operand, pop-pop-apply-push-on-operator without any parsing distractions.
4 free resourcesSolve → - #152mediumfrequently asked
152. Maximum Product Subarray
Given an integer array, find the contiguous subarray with the largest product. LinkedIn asks this on the DP round because negatives turn Kadane on its head — you need to track BOTH min and max running products to handle sign flips.
4 free resourcesSolve → - #208mediumcompany favorite
208. Implement Trie (Prefix Tree)
Implement a Trie with insert, search, and startsWith. LinkedIn asks this because it's the foundation of their typeahead and autocomplete features — they want a clean node-with-children-map structure and the isEnd boolean.
4 free resourcesSolve → - #230mediumfrequently asked
230. Kth Smallest Element in a BST
Given a BST and integer k, return the kth smallest value. LinkedIn asks this on the BST round — they want the iterative inorder with early-stop, AND the follow-up answer if the tree is modified frequently (augment with subtree size).
4 free resourcesSolve → - #243easycompany favorite
243. Shortest Word Distance
Given an array of strings and two words, return the shortest distance between two indices where these words appear. LinkedIn asks this as the entry point to their signature Shortest Word Distance family — they want the single-pass O(n) solution with two index trackers.
4 free resourcesSolve → - #244mediumcompany favorite
244. Shortest Word Distance II
Same as Shortest Word Distance but with many queries. LinkedIn asks this to test whether you can amortize cost across queries — pre-bucket indices by word, then run a merge-like two-pointer walk per query.
4 free resourcesSolve → - #245mediumcompany favorite
245. Shortest Word Distance III
Same as Shortest Word Distance but word1 might equal word2 — meaning you need the closest pair of occurrences of the same word. LinkedIn asks this as the culminating follow-up to the series, testing whether you spot the asymmetric case before writing code.
4 free resourcesSolve → - #277mediumcompany favorite
277. Find the Celebrity
Given n people and a knows(a, b) API, find THE celebrity — someone known by everyone but who knows no one. LinkedIn asks this because it's the canonical 'elimination by partial order' problem — n calls find a candidate, n calls verify.
4 free resourcesSolve → - #297hardfrequently asked
297. Serialize and Deserialize Binary Tree
Design an algorithm that serializes a binary tree to a string and back. LinkedIn asks this on senior loops because the preorder-with-nulls + queue-based recursive parse is the cleanest pair you can write under interview pressure.
4 free resourcesSolve → - #339mediumcompany favorite
339. Nested List Weight Sum
Given a nested list of integers, return the sum of all integers weighted by their depth. LinkedIn asks this on the recursion round — they want clean DFS or BFS code that uses the provided NestedInteger interface without unpacking it eagerly.
4 free resourcesSolve → - #364mediumcompany favorite
364. Nested List Weight Sum II
Same as Nested List Weight Sum but the weighting is INVERTED — deepest leaves are weight 1 and the root is the heaviest. LinkedIn asks this as the trickier follow-up; the clean solution is a 'cumulative sum at each depth' trick that avoids two passes.
4 free resourcesSolve → - #366mediumcompany favorite
366. Find Leaves of Binary Tree
Given a binary tree, collect its leaves and remove them, then repeat. Return the order of removals. LinkedIn asks this because the elegant solution isn't iterative removal — it's a 'height from leaf' DFS that buckets each node by its eventual removal layer.
4 free resourcesSolve → - #605easyfrequently asked
605. Can Place Flowers
Given a flowerbed (0s and 1s, no two adjacent 1s) and a number n, return whether you can plant n new flowers without breaking the no-adjacent rule. LinkedIn asks this as the greedy warm-up — they want clean per-cell checks with no boundary bugs.
4 free resourcesSolve → - #797mediumfrequently asked
797. All Paths From Source to Target
Given a DAG with n nodes labeled 0 to n-1, find all possible paths from node 0 to node n-1. LinkedIn asks this on the graph traversal round — they want clean DFS with path tracking, no DP needed (because it's a DAG and paths could be exponential).
4 free resourcesSolve →
Related interview-prep guides
HackerRank Tech Interview Guide 2026: What It Tests, How It Tracks You, and the Modern Setup
HackerRank is still the volume leader in first-round technical screens for 2026 tech hiring. A browser-sandboxed coding environment that logs every keystroke, paste event, and tab-focus change inside its own tab. This guide covers what it tests, the boundary of what it can and cannot detect, and how a modern desktop setup pairs with a HackerRank session without leaking into the screen-share.
Karat Technical Interview Guide 2026: How the Third-Party Loop Actually Works
Karat is technical-interview-as-a-service. Karat-employed engineers run the technical loop for the hiring company in Karat's own recorded video and coding environment. The dynamic is different from an in-house interview: the interviewer is a contractor, not a future teammate, the rubric is fixed, the session is recorded for asynchronous review, and the hiring team's engineers watch the playback a day later. This guide is the practical map of how that loop works in 2026 and how a modern desktop setup runs alongside it.
Microsoft Teams for Tech Interviews in 2026: The Complete Candidate Guide
Microsoft Teams is the default interview surface at any company running Office 365: Fortune 500, finance, healthcare, government tech, legacy enterprise. The recording-and-transcript reality changes the threat model versus Zoom, but the OS-level boundary that protects a modern desktop AI setup is the same. This is the candidate-side guide to running a Teams interview in 2026.