Microsoft Interview Guide

Microsoft’s loops are the most collaborative of the big five. Interviewers routinely work the problem with you, offer hints early, and treat a hint taken well as a positive signal rather than a deduction. Candidates coming from a Google-shaped mental model often over-perform the stoic lone-genius act and score worse for it.

The bar leans toward clean, production-shaped code rather than exotic algorithms: sensible names, handled edge cases, a function you would be willing to review. Expect mediums, expect to be asked about testing, and expect the conversation to matter.

Crowd-reported, not published

No company publishes its question bank. The round structure and the bar below are corroborated by recruiters and candidates; the problem list is crowd-reported, stale, and biased toward whoever posted it. Treat the pattern families as the durable signal — which families a company leans on changes far more slowly than which individual problems it uses.

The most conversational of the big five, and the most likely to care about code quality rather than just correctness. Expect discussion of edge cases, naming and testability.

The loop

Recruiter + technical screen 1× · 60 min One medium, often with a design flavour

Onsite 3-4× · 60 min Coding plus practical design and past-project depth

As-appropriate ("AA") round 1× · 60 min Senior interviewer, broad and decisive

The bar

Write code you would submit in a pull request: named variables, guarded edges, and a stated test plan. "How would you test this?" is asked far more often here than elsewhere.

What they lean on

This is the durable part. Which pattern families a company favours is far more stable than which individual problems it uses, so prepare in this order:

1. linked-lists
2. binary-trees-and-bst
3. strings
4. design-with-stacks-and-queues
5. matrix-and-grid-manipulation
6. one-dimensional-dp

Quirks worth knowing

• Expect to be asked to walk your own code through a test case line by line.
• Real-world framing is common ("this array is a log of requests"), so restate the problem in plain terms first.
• Deep questions about your own past projects are standard. Have one project you can defend at architecture level.

Reported problems

Read this before you trust the list. No company publishes its question bank. Every "asked at Microsoft" list in circulation — including this one — is aggregated from candidate reports, is months to years out of date, and over-represents whoever bothered to post. Use it to choose an order to practise in. Do not use it as a question bank, and do not assume a problem missing from it is safe to skip.

24 problems

7 easy16 medium1 hard

AllEasyMediumHard

Unsolved onlyExportImportReset

• ○1Two SumeasyNeetCode 150Blind 75LeetCode Top Interview 150
• ○20Valid ParentheseseasyNeetCode 150Blind 75LeetCode Top Interview 150
• ○121Best Time to Buy and Sell StockeasyNeetCode 150Blind 75LeetCode Top Interview 150
• ○206Reverse Linked ListeasyNeetCode 150Blind 75
• ○21Merge Two Sorted ListseasyNeetCode 150Blind 75LeetCode Top Interview 150
• ○141Linked List CycleeasyNeetCode 150Blind 75LeetCode Top Interview 150
• ○217Contains DuplicateeasyNeetCode 150Blind 75
• ○3Longest Substring Without Repeating CharactersmediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○153SummediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○53Maximum SubarraymediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○146LRU CachemediumNeetCode 150LeetCode Top Interview 150
• ○200Number of IslandsmediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○33Search in Rotated Sorted ArraymediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○72Edit DistancemediumNeetCode 150LeetCode Top Interview 150
• ○75Sort Colorsmedium
• ○102Binary Tree Level Order TraversalmediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○236Lowest Common Ancestor of a Binary TreemediumLeetCode Top Interview 150
• ○238Product of Array Except SelfmediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○300Longest Increasing SubsequencemediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○322Coin ChangemediumNeetCode 150Blind 75LeetCode Top Interview 150
• ○1143Longest Common SubsequencemediumNeetCode 150
• ○138Copy List with Random PointermediumNeetCode 150LeetCode Top Interview 150
• ○994Rotting OrangesmediumNeetCode 150
• ○4Median of Two Sorted ArrayshardNeetCode 150LeetCode Top Interview 150

Pitfalls

• Refusing hints. A hint accepted and built on is a positive signal here. Silently struggling for ten minutes is not.
• Writing contest-style code. Single-letter names and no edge-case handling read badly when the bar is “would I approve this pull request?”
• Not asking about testing. Offer the test categories unprompted — it is closer to the day job than the algorithm is, and it is explicitly valued.
• Treating it as easier because it is friendlier. The collaboration is genuine; the bar is not lower. Clean beats clever, and clean is harder to fake.

Interview follow-ups

They ask	What they’re checking	The answer
“Would you like a hint?”	Whether you collaborate	Yes — and say what you were stuck on first. Naming the stuck point turns the hint into a dialogue rather than a rescue.
“How would you test this?”	Engineering instinct	Categories rather than examples: empty, one element, duplicates, maximum size, and the specific boundary the algorithm pivots on.
“What would you extract if this were production code?”	Whether you know interview code is not production code	Input validation, a named helper for the core transform, magic numbers as constants — while saying you deliberately skipped them for the 45-minute version.
“Can you make this more readable?”	That clean is the actual bar	Take it seriously: extract the condition into a named predicate, split the loop body, rename i/j to what they index.

Recall card

• Collaborative by design. Hints are offered early and taking one well is a positive signal.
• The bar is clean, reviewable code — names, edge cases, a function you would approve. Clean beats clever.
• Offer the test categories unprompted. It is explicitly valued and closest to the day job.
• Friendly is not easier. Say what you are stuck on rather than going quiet.
• Know what you would extract for production — and say you skipped it on purpose.