Jobra
🐞Step 1 of Jobra:AI Debugging Rounds
DSA+AIPreparing you for both.
JOBRA · AI DEBUG ROUND (PREVIEW)sum.cpp
// buggy snippet
int total(vector<int>& a, int n) {
    int sum = 0;
    for (int i = 0; i <= n; i++) { // 🐞        sum += a[i];
    }
    return sum;
}
A real codebase. A real bug. You + AI.

Jobra drops you into a broken repo. Find it, fix it, ship it. AI helps you debug like on the job.

Prompt:Correct the loop condition
Tap 'Let Jobra fix it' below to simulate AI debugging...

Samsung Software Engineer Interview Questions & Rounds

32 min read

How the Samsung software engineer interview works in India: the 3-hour coding test, technical and HR rounds, with questions from real candidate experiences.

In this article
On this page

How to Crack Samsung: The Software Engineer Interview, From Real Experiences

Last updated November 20, 2026 · 27 min read · For interns and new grads at Samsung's India R&D centres (SRI-Bangalore, SRI-Noida, SRI-Delhi, SSIR)

Built by the Codinzhub team. This guide is built from Samsung software engineer interview experiences that candidates shared, including posts and coding-test questions on Codinzhub. The process changes by year, centre and campus, so treat this as a map, not a promise.

The 30-second version. Most Samsung software engineer interview reports come from campus drives for Samsung's R&D centres in India. After a CGPA shortlist (cut-offs of 7 to 8.5), most candidates sat a 3-hour coding test with one problem, and in 8 of 10 such reports only those who passed all the hidden test cases got an interview. Some drives used a 60-minute, 3-problem round or MCQs instead. Then came usually one technical interview, 30 to 60 minutes in most reports that give a length, mixing a few DSA problems with OS, OOP and your projects, and finally an HR round that can still reject you. In the reports that give timing, interviews came the same day as the test. The habit that matters most: test your own edge cases before every capped submission.

What this guide is, and what it is not

This guide is built from 32 candidate experiences of Samsung's software engineer interviews in India, mostly from 2023 to 2026, at the Samsung R&D institutes in Bangalore, Noida and Delhi, Samsung Semiconductor India Research and other Samsung R&D centres. Most reports are campus hiring, and 17 are internships.

It is:

  • A map of the process, stage by stage, as candidates described it
  • The question patterns that keep coming back, and the follow-ups to expect
  • A six-week plan and practice problems matched to those patterns

It is not:

  • A list of leaked questions or a promise of what you will be asked
  • An official description of Samsung's process
  • An offer rate. Outcomes in candidate posts are self-reported and are not a statistical sample.

Important things to know before you start

These Samsung-specific facts change how you should prepare.

The coding test is all or nothing. Most reports (10 of 17) describe one problem in 3 hours on Samsung's own test software. In 8 of those 10, you had to pass all the hidden test cases (10 or 50) to get an interview, with submissions capped at 5 or 10. Write and run your own edge cases before each submission.

Check the library rules on the day. Two older reports (2018 and 2020) said C++ STL or many library functions were not allowed in the test. No 2023 to 2026 report mentions this rule, but be able to write a queue, a sort and a hash map yourself in case.

The interview follows your resume. Projects came up in 18 of 28 interview reports and core CS subjects (OS, OOP, DBMS) in 17 of 28. Interviewers asked about whatever you listed, from Linux and DevOps to RAG. List only what you can defend.

Round-by-round breakdown

Candidates described five stages: application, an academic shortlist, a coding test, one or two technical interviews followed by an HR round, and the result.

Stage 1: Application, where candidates applied

Almost all of this data is campus hiring in India: 26 of the 27 records that name a channel. The drives covered the Bangalore, Noida and Delhi R&D institutes (SRI-B, SRI-N, SRI-D), Samsung Semiconductor India Research (SSIR) and unnamed Samsung R&D centres.

ChannelWhat candidates said
Campus driveUsed for 2-month summer interns, 6-month interns, research interns and full-time engineers. One 6-month SDE intern applied through the college portal, and the resume and CGPA were reviewed before the test invite. One 2024 full-time drive was open only to women students and was held at the SRI-D office in Noida.
Off campusOne report: a C++ software engineer role on device drivers, with a 30-question C++ test, two technical rounds and HR.

What helped. Two selected SDE I candidates credited starting preparation about a year before placements, and one of them also took regular coding contests. One 2023 candidate advised solving previous years' Samsung coding-test problems.

Stage 2: Shortlisting, how candidates moved forward

Shortlisting happened twice on campus: on academics before the test, then on the test result before interviews.

  1. Academic cut-off (7 records). CGPA cut-offs ranged from 7 to 8.5. Two drives also needed 60% in class 10 and 12. Branch rules varied: one campus took only B.Tech CS (not IT), one 2025 drive took only CSE and ECE above 8.0, and one SRI-B intern drive took only circuit branches above 7.5.
  2. Resume shortlist instead of a test (1 record). A 2023 SRI-N intern drive skipped the coding test. It shortlisted 27 candidates on their resumes (CGPA 8.5 and above) and went straight to interviews.
  3. The coding-test cut. In 8 of the 10 three-hour test reports, only candidates who passed all the hidden test cases got an interview. Reported numbers: about 30 qualifiers at one 6-month intern drive; 15 plus 5 on standby out of 135 registrations at another; about 50 interviewed at an SRI-B intern drive; and 6 software candidates out of 250 at SSIR.
  4. An extra pen-and-paper round (1 record). At a 2025 SRI-N drive, about 40 candidates cleared the test, so Samsung ran a written round that cut the list to about 10. The candidate said this round is held only when too many people qualify.

Stage 3: The coding test

Most candidates sat a coding test before any interview. The format is what sets Samsung apart, and most reports describe the same one across SRI-B, SRI-N and SRI-D.

Online coding test

How it works. In 10 of the 17 reports that describe the test, it was one problem in 3 hours, with hidden test cases. In 8 of those 10, you had to pass all the test cases to reach the interview; partial credit did not count. Reports gave either 10 test cases (3 reports) or 50 (4 reports). You could compile and run as often as you liked, but submissions were capped: 5 in two reports, 10 in two others. Four reports describe Samsung's own test software (Windows-only in one, flagging remote-access tools in another), and four tests were invigilated in person. No report mentions AI-tool rules.

At a glance

  • Length: 3 hours in 10 of 17 reports. Other formats that give a length ran 60 to 90 minutes.
  • Questions: 1 coding problem in the 3-hour format; all hidden tests must pass (8 of 10).
  • What was asked: Graph and grid problems, dynamic programming and binary search over a formula; several problems are known only by name. A 2025 candidate called it a competitive-programming problem of moderate rating rather than a typical interview problem.
  • Most common: No single repeated question. Burst Balloons and a find-n-in-a-monotonic-equation problem each appeared in 2 reports.

Other formats in the data. The exceptions follow centre and track: SSIR in 2023 gave 3 problems in 60 minutes; one research-intern test had 3 problems; one 2025 SRI-N drive used 50 MCQs on aptitude and CS basics in 1 hour; a 2026 AI developer test had 70 AI MCQs and 2 prompt-engineering questions; and an off-campus C++ role had 30 C++ objective questions in 60 minutes.

QuestionWhat it testsLevel
Burst Balloons: maximise coins by choosing the burst orderInterval DPNew grad
Find n where a monotonic equation f(n) equals a given valueBinary search over a monotonic functionSWE
Graph problem built around cycle detectionCycle detection in a graphIntern
Flip whole columns exactly k times to maximise all-ones rowsGrouping identical rows, parity of leftover flipsIntern
Kim and his Refrigerators (problem named; statement not given in the report)Not reportedNew grad
Count substrings of s that read as a power of 2 (pen-and-paper round)Not reportedSWE
Grid DP with backtracking, similar to Cherry Pickup IIMulti-agent grid DPIntern
Subtract two numbers stored as linked listsDigit-by-digit arithmetic on listsIntern

Coding-test questions shared on Codinzhub. These pages show only the title and a short description publicly, so the topic is Not reported where the technique is not visible.

QuestionDriveTopic
Spaceship Coin CollectorSamsung R&D, college not shownNot reported (grid problem)
Doctor's Location ProbabilitySamsung R&D, college not shownGraphs (directed graph)
Research Center LocationSamsung R&D, college not shownNot reported
N machines with unique machine IDsSamsung R&D Noida, IIT (ISM) DhanbadNot reported

See all Samsung coding-test questions in the question bank.

Stage 4: The interview loop

The loop is short. Of the 20 reports that give a round count, 15 had one technical interview and 5 had two. Most reports (16 of 28) then describe an HR round, sometimes on the same day as the test. Interviews were in person in 5 reports and online (MS Teams or a shared editor) in 2. No record describes a separate system design round.

Technical interview: DSA, core CS and your projects

At a glance

  • Length: 15 to 120 minutes; 30 to 60 minutes in 5 of the 7 reports that give a length.
  • Questions: One technical interview in 15 of 20 reports; often 2 to 4 coding problems plus rapid concept questions.
  • What was asked: DSA plus core CS (OS, OOP, DBMS) in 17 of 28 reports. Projects came up in 18 of 28. Research-intern and AI roles added ML questions (4 reports).
  • Most common: What is context switching? (4 reports, all from one 2026 IIT Patna research-intern drive).

The interview follows your resume. A candidate who listed Linux got a bash-scripting task, one who mentioned DevOps got Docker and Jenkins questions, and an AI developer with RAG on the resume had to explain how it works. Research interns met ML interviewers who asked about overfitting and validation sets; one cut a candidate off for talking about development skills instead of AI. Four reports say the interviewer wanted the approach or pseudo code, not running code. Difficulty varied by interviewer: one 2023 SRI-N intern got an LRU cache and then a LeetCode-hard segment-tree problem, while one 2025 SRI-N drive asked only DSA and design (LRU, LFU) with no core subjects at all.

Reported coding problems

QuestionFollow-upLevel
Empty cell farthest from its nearest city on a gridAsked open-ended first: where to put a lab so the cities stay safeIntern
Word Ladder: fewest one-letter changes from source to targetNot reportedIntern
Decode a nested k[encoded] stringNot reportedIntern
Longest consecutive sequence in O(n)Not reportedIntern
Merge overlapping intervalsNot reportedIntern
Largest subarray with sum 0Brute force O(n^2) first, then O(n) with a hash mapNew grad
Design an LRU cache, then count cache hits and missesDesign an LFU cache (one 2025 report)Intern, SWE
First 1 in a hidden 0/1 array with two strikes and 150 readsNot reportedIntern

Reported core CS, ML and design questions

QuestionWhat it testsLevel
What is context switching, and exactly what state is saved and restored?OS fundamentalsIntern
Explain the four pillars of OOP with real examplesOOP fundamentalsSWE, New grad
How would you fix overfitting?ML fundamentals; follow-up removed the option of changing the architectureIntern
Deadlock: conditions, how it occurs, and ways to prevent itOS synchronisationIntern, New grad
Mutex, locks and semaphoresOS synchronisationIntern, New grad
Virtual inheritance and the diamond problemC++ object modelSWE, New grad
Two threads print 1 to 10 in order from odd and even vectorsThread synchronisationSWE
Design a URL shortener (concept level)High-level design basicsIntern

Puzzles came up too, such as a bee flying between two approaching trains.

HR round

At a glance

  • Length: 25 minutes in one report; the others did not say.
  • Questions: Background, motivation and your projects again; puzzles for some research interns.
  • Themes: Family and education, why Samsung, projects, commitment to joining, career plans.
  • Most common: Tell me about your family background (7 of the 16 reports with an HR round).

Two reports called the HR round relaxed or quick, but it is not a formality. One 6-month SDE intern cleared the technical round, then was rejected in HR and blamed over-explaining. At an SSIR drive and a 2023 SRI-N intern drive, 3 candidates reached HR and 2 were selected.

Reported questions.

  • Tell me about your family background (7 reports)
  • Why Samsung? (5 reports)
  • Walk me through your projects again: why you built them, solo or group (5 reports)
  • Would you still join if there is no pre-placement offer (PPO)? (3 research-intern reports)
  • Cut a cream-topped cake into 8 equal pieces with 3 cuts (3 reports; one interviewer rejected the horizontal cut because only the top has cream)
  • How did your technical rounds go? (3 reports)
  • JEE rank, category and any drop year (2 reports)
  • Tell me about a conflict with your mentor in a previous internship
  • What differentiates you from others? Then: plans for higher studies, and any location constraints

See all HR questions in the question bank.

Stage 5: Result and offer

Result. Campus drives moved fast. One 6-month intern heard about the shortlist about an hour after the test and had HR about 20 minutes after the technical round. A 2024 full-time drive held interviews after lunch on the day of the test. At one SRI-B intern drive, interviews went in CGPA order, highest first. No record gives an end-to-end timeline in weeks.

How many were picked. Reported funnels: 135 registrations, 15 interviews plus 5 on standby, 5 selected; about 40 test qualifiers, about 10 after the pen-and-paper round, 5 selected; about 50 interviewed and about 30 selected at an SRI-B intern drive; 12 interviewed and 10 selected for research internships. These are single drives, not offer rates.

Offer and conversion. One 2024 full-time candidate had the CTC breakdown explained in the HR round. SSIR offered a 6-month internship with performance-based conversion to a full-time role. For IIT Patna research interns in 2026, one interviewer said Samsung usually does not give interns a PPO, and three HR rounds asked whether the candidate would join anyway.

Samsung experiences shared on Codinzhub

These are Samsung R&D internship experiences that candidates posted on Codinzhub. Only a short description is public, so the summaries are brief.

CollegeRole and resultWhat was asked
IIT KharagpurSamsung Research Institute intern, selected (2026-27 season)Puzzles on sand clocks; topological sort
IIT (ISM) DhanbadSamsung R&D Noida 2-month intern, selectedRound 1 focused on DSA and core computer science
IIT PatnaSamsung R&D intern, outcome not reportedTwo questions, the first on counting connected components

The Codinzhub Next-Round Playbook

Candidates who moved forward treated the test as all-or-nothing and the interviews as a conversation about their own resume.

To get past the shortlist

  • Check the drive's CGPA and branch rule before you plan around it. Cut-offs in the data ran from 7 to 8.5, and some drives excluded IT or non-circuit branches.
  • Make your resume interview-ready. One 2023 SRI-N intern drive shortlisted on resumes alone and skipped the test.

To get past the coding test

  • Aim for all test cases, not most of them. In 8 of the 10 three-hour test reports, one failing hidden case meant no interview.
  • Write your own edge cases and run them freely, but submit carefully: two reports capped submissions at 5, two at 10.
  • Practise in plain C++ or Java with your own helpers; the test runs on Samsung's own software, not a familiar judge.

To get past the technical interview

  • Lead your introduction with your strongest area. Interviewers followed it: Linux led to a bash script, DevOps to Docker and Jenkins, an RAG project to how RAG works.
  • Give the brute force, then improve it out loud. One SDE I candidate went from O(n^2) to O(n) on the zero-sum subarray, and the interviewer was satisfied.
  • Be ready to explain without code. Four reports say only the approach or pseudo code was wanted.

To get past the HR round

  • Keep answers short. One 6-month intern cleared the technical round and was rejected in HR, and blamed over-explaining.
  • Prepare a clear "why Samsung" and an honest answer on joining without a PPO; both came up repeatedly.
  • Be ready to describe your projects again in plain words. Five HR rounds went back to them.

To get through the result

  • Keep your phone on during the drive: one shortlist arrived about an hour after the test.
  • Ask about the internship length and conversion terms in HR. Reports mention 2-month and 6-month internships, and a CTC explanation in one full-time HR round.

How the loop differs by level

The core path, a coding test then a short loop, is the same for interns and full-time hires. Research and AI roles add ML, and two of three full-time reports had a second technical round.

LevelWhat candidates describe
Intern (17 records)2-month summer, 6-month and research internships. Usually the 3-hour test (or a 60-minute, 3-problem round at SSIR), then one technical interview (9 of the 10 intern reports that give a round count) and HR. Research interns at IIT Patna in 2026 got grid BFS problems, OS questions and ML questions on overfitting and validation sets. No intern had a system design round.
New grad (3 records)SDE I and graduate engineer trainee (GET) roles, all through the 3-hour test. Two SDE I reports had two technical rounds, then HR; the GET report had one. Questions included the largest zero-sum subarray, top and bottom view of a binary tree, OOP pillars, deadlocks and semaphores. No system design round.
Level not stated (12 records)Campus software engineers, an AI developer (70 AI MCQs, then a 30-minute online interview on Kadane's algorithm and RAG), an embedded-software candidate (MCQs, structure padding, a separate business call round) and one off-campus C++ device-driver engineer (C++ test, two technical rounds on C++, OS and threads).

There are no experienced-hire (SDE 2 and above) reports in this data, so this guide does not cover lateral loops.

How the centres differ. The 3-hour single-problem test appears at SRI-B, SRI-N and SRI-D, so this guide treats them as one process. The differences are in the details:

CentreWhat the data shows
SRI-Bangalore (2 records)3-hour test with 50 hidden tests on Samsung's Windows-only software; one short interview, ordered by CGPA.
SRI-Noida (5 records)Most varied. One 2025 drive used the 3-hour test plus a pen-and-paper cut and asked only DSA and LRU/LFU design; another 2025 report had a 50-MCQ written test, then DSA and DevOps questions; a 2023 intern drive skipped the test entirely.
SRI-Delhi (2 records)3-hour test with 10 hidden tests and 5 submissions. A 6-month intern faced 3 or 4 graph questions with Dijkstra and Floyd-Warshall follow-ups; a 2024 GET drive covered OOP and DSA, then HR.
SSIR (1 record)60-minute, 3-problem coding round; a 6-month internship with conversion to full time; a 1 hour 25 minute online technical interview, then HR.

Most-asked topics

Across the 31 reports that gave coding-test or interview questions, graphs and grids led. Dynamic programming was less common than its reputation suggests: 5 reports.

TopicWhat it meansSeen inReported examples
Grids and graphs (BFS, DFS)Cells or nodes connected by moves or edges; explore level by level or deep first16 of 31 reportsEmpty cell farthest from its nearest city; Word Ladder; graph cycle detection
Arrays, hashing and prefix sumsSingle-pass scans, hash sets and maps, running totals15 of 31 reportsLongest consecutive sequence; largest zero-sum subarray; merge intervals
Core CS: OS, OOP and DBMSProcesses, threads, locks, memory, OOP pillars, SQL11 of 31 reports (17 of 28 interviews counting topic-only mentions)Context switching; deadlock; four pillars of OOP
Linked lists, stacks and queuesPointer manipulation and LIFO or FIFO structures9 of 31 reportsDecode a nested string; reverse a linked list; subtract two list numbers
TreesBinary-tree traversals, views and path sums7 of 31 reportsBinary tree maximum path sum; top and bottom view; lowest common ancestor
Binary search and mathsSearching a monotonic answer, number tricks7 of 31 reportsFind n in a monotonic equation; repeated integer square root; two-strike search
Dynamic programmingBuilding answers from overlapping subproblems5 of 31 reportsBurst Balloons; Cherry Pickup II style grid DP; longest common subsequence
Machine learningOverfitting, validation, training curves4 of 31 reports (research and AI roles)Fix overfitting; early stopping; how RAG works

The question patterns that keep coming back

Most questions resolved into a handful of patterns. Learn to spot the pattern under the story.

1. Grids and graphs, mostly BFS

The largest pattern (16 of 31 reports), in both the test and the interview. The disguise is usually a grid of cities or land, or words that differ by one letter. Choose the technique by the question: fewest steps or nearest distance means BFS, and distance from many sources at once means multi-source BFS (put every source in the queue at the start). The most repeated question was the empty cell farthest from its nearest city (3 reports).

Reported examples. Empty cell farthest from its nearest city; distance to the nearest 0; Word Ladder; empty cell minimising total distance to all cities; counting connected components.

2. Arrays, hashing and prefix sums

Seen in 15 of 31 reports. If the question asks for a subarray with a target sum, store prefix sums in a hash map. If it asks for runs or membership in O(n), use a hash set. If it involves overlapping ranges, sort by start first.

Reported examples. Longest consecutive sequence in O(n); largest subarray with sum 0; merge overlapping intervals; vowels in query ranges; 3Sum and 4Sum.

3. Core CS: OS, OOP and DBMS

Specific core CS questions appear in 11 of 31 reports, and 17 of 28 interview reports touched these subjects. Context switching was the most repeated (4 reports, all from one 2026 drive). Answer with the mechanism, not the definition: what state is saved, who saves it, and what it costs. For synchronisation, be able to write a small two-thread example with a mutex.

Reported examples. Context switching; multiprocessing; deadlock conditions and prevention; mutex, locks and semaphores; four pillars of OOP; SQL vs NoSQL.

4. Linked lists, stacks and queues

Seen in 9 of 31 reports. Use two pointers (slow and fast) for middles and cycles, and a stack when brackets or nesting appear.

Reported examples. Decode a nested k[encoded] string; reverse a linked list; detect a cycle; reverse a stack using recursion.

5. Trees

Seen in 7 of 31 reports. Use level-order BFS for views and level sums, and post-order DFS when a node needs answers from both children, as in path sums.

Reported examples. Binary tree maximum path sum; top and bottom view; lowest common ancestor; sum of alternate levels.

6. Binary search and maths

Seen in 7 of 31 reports. If the answer is a number and a check gets easier as it grows, binary search on the answer. The two-strike search is the two-egg drop problem: with q probes and 2 strikes you can cover q(q+1)/2 positions.

Reported examples. Find n where a monotonic equation equals a value; repeated integer square root in O(log n); first 1 in a hidden 0/1 array with two strikes; A+B using only increments.

7. Dynamic programming

Seen in 5 of 31 reports, but it included two coding-test problems that candidates called hard. When the order of choices matters, as in Burst Balloons, think interval DP over the last choice made. Show recursion, then memoisation.

Reported examples. Burst Balloons; grid DP similar to Cherry Pickup II; longest common subsequence; maximum subarray sum.

Preparing for the HR round

The HR round can still end your run (see Stage 4). It is about you and your reasons, so prepare honest, short answers rather than long stories.

Build your answer bank. Prepare short answers, each under a minute, for the themes candidates were asked about:

ThemeReported prompt (paraphrased)Follow-up to expect
BackgroundTell me about your family backgroundJEE rank, category and any drop year (2 reports)
MotivationWhy Samsung?Would you still join if there is no PPO?
ProjectsWalk me through your projects againWhy you built it, and was it solo or group?
Self-assessmentHow did your technical rounds go?Which questions you got and how you solved them
ConflictA conflict with your mentor in a previous internshipHow you handled the challenges
FutureWhere do you see yourself in 5 years?Plans for higher studies; any location constraints
PersonalitySomething about you that is not on your resumeWhy only technical skills in your intro?

Keep it brief. One candidate felt that over-explaining cost them the round. Answer the question, give one example, and stop.

Be honest about other options. Two research interns were asked which other companies they had applied to, and three whether they would join without a PPO.

Have a puzzle ready. Research-intern HR rounds included the cake-cutting and climbing-monkey puzzles. Reason out loud and check constraints, such as the cream only being on top.

A six-week plan

The plan follows the patterns above, most common first, and ends with mocks in the real formats: a 3-hour single-problem test and a 30- to 60-minute interview. Plan on two to three focused hours a day.

  1. Week 1: Grids and graphs. BFS, multi-source BFS, DFS, cycle detection, connected components, shortest paths. Habit: draw the grid or graph and trace the queue by hand before coding.
  2. Week 2: Arrays, hashing and prefix sums. Hash sets and maps, prefix sums, intervals, two pointers (3Sum, 4Sum). Habit: say the brute force and its complexity first, then improve it.
  3. Week 3: Core CS. Context switching, processes vs threads, deadlock, mutexes and semaphores, scheduling, OOP pillars and virtual inheritance, SQL basics. Habit: explain each idea with one concrete example in two sentences.
  4. Week 4: Linked lists, stacks, queues and trees. Reversal, cycles, nested decoding, tree views, path sums, LCA. Habit: write the full code by hand once a day, without autocomplete.
  5. Week 5: Binary search, maths and DP. Binary search on the answer, egg-drop reasoning, interval DP (Burst Balloons), grid DP, LCS. Research and AI candidates add overfitting, validation sets and early stopping, plus one deep dive on your own ML project.
  6. Week 6: Mocks. Two full 3-hour single-problem mocks with your own edge cases and a 5-submission cap, plus four 45-minute interview mocks that start from your resume and end with OS questions. Rehearse your HR answers out loud.

Interns can go lighter on hard DP; full-time candidates should expect a possible second technical round. No level had a system design round, but practise explaining an LRU cache and your project's architecture.

Only two weeks? Spend four days on grids and graphs, three on arrays and hashing, two on core CS, two on linked lists and trees, one on binary search and DP, and two on a 3-hour mock plus interview and HR rehearsal.

Practice problems that match these experiences

These are well-known practice problems that match the patterns above. They are not claims about Samsung's question bank unless a row says the problem was asked as-is.

#Practice problemPatternSeen in these experiences as
1As Far from Land as PossibleMulti-source BFSAsked as-is in three research-intern reports (the lab-placement grid)
2Word LadderBFS on an implicit graphAsked as-is in two reports
3Burst BalloonsInterval DPAsked as-is as the 3-hour test problem in two reports
4Find element position in given monotonic sequenceBinary search on a monotonic functionAsked as-is as the 3-hour test problem in one report; a second report matches it
5Largest Subarray with 0 SumPrefix sums with a hash mapAsked as-is in two SDE I reports
6Longest Consecutive SequenceHash setAsked as-is in two reports
7Decode StringStackAsked as-is in two reports
8LRU CacheHash map plus doubly linked listLRU cache with hit and miss counts
9Top View of Binary TreeLevel-order with horizontal distanceAsked as-is in one SDE I report; another got the bottom view
10Egg Drop With 2 Eggs and N FloorsTwo-egg drop reasoningThe two-strike search on a hidden 0/1 array
11Print Zero Even OddThread hand-off with locksTwo threads printing 1 to 10 in order

How to handle a question you have never seen

Interviews here are a few short problems plus follow-ups, so a repeatable method matters more than memorised solutions.

  1. Repeat the problem and one example. Open-ended prompts happen: one research intern had to define what "safe" meant before the lab-placement grid became a BFS problem.
  2. Say the brute force and its complexity. It gives the interviewer something to push on, and one SDE I candidate moved from O(n^2) to O(n) this way.
  3. Name the bottleneck. Repeated lookups suggest a hash map; repeated ranges suggest prefix sums; nearest or fewest steps on a grid suggests BFS; a number answer with a monotonic check suggests binary search.
  4. Say when you are stuck. One SRI-B intern could not improve a string-compression answer and the interview ended soon after; a research intern who kept explaining a BFS idea was told it was close to optimal. Talk through what you have.
  5. Write pseudo code in clear pieces. Four reports say the approach or pseudo code was enough, so name your helper steps and keep the structure clean.
  6. Dry-run one case and fix the wrong line. Trace your example by hand, including an edge case such as no empty cells or an empty list.

Why candidates were rejected

Seven records end in rejection, and 4 of them give a reason. All four are the candidates' own view, not recruiter feedback.

  • Over-explaining in HR. One 6-month SDE intern cleared the technical round but was rejected in HR, and felt that long answers were the reason.
  • Stopping short of the optimal solution. One 2023 SRI-N intern solved an LRU cache quickly but could not reach the segment-tree solution for a hard follow-up problem, and was rejected after that round.
  • A weak answer to an open-ended question. One SRI-B intern gave a frequency-based answer to a string-compression question; the interviewer seemed unsatisfied and the interview ended shortly after.
  • Profile fit for a specialised role. One embedded-software candidate felt their non-electrical background counted against them.

The other three rejected records gave no reason.

Selected vs. rejected: do this, not that

No selected candidate described a perfect interview; what repeated were a few habits.

Do thisNot thatWhat the reports show
Pass everything: test edge cases before each capped submissionSubmit and hope: treat the test like a partial-credit contestIn 8 of 10 three-hour test reports, failing one hidden case ended the process.
Answer briefly: one point and one example in HROver-explain: long answers to simple questionsOne intern cleared the technical round and felt over-explaining cost them HR.
Own your resume: know every skill and project you listList what you cannot defend: buzzwords you cannot explainInterviewers followed the resume in 18 of 28 reports; one selected SDE I candidate warned against listing skills you do not know.
Improve out loud: brute force, then the optimisationFreeze on the optimal: silence when the best approach is not obviousOne selected SDE I candidate went from O(n^2) to O(n); one rejected intern could not reach the optimal solution.

If you remember only five things

  1. The process is short. A CGPA shortlist, a coding test, usually one technical interview, then an HR round; on campus, the reports that give timing had interviews the same day as the test.
  2. The test is one problem in 3 hours. In most reports you had to pass all the hidden test cases to reach the interview.
  3. Graphs and grids lead the questions. BFS on grids, arrays with hashing, and core CS (context switching, deadlock, OOP) are the core.
  4. Your resume sets the questions. Interviewers followed your projects and listed skills, and four reports wanted only the approach or pseudo code.
  5. HR is not a formality. Keep answers short and know why you want Samsung, even without a PPO.

FAQ

Is there a coding test at every drive?

No. The 3-hour single-problem test appears in 10 of 17 reports, but SSIR used a 60-minute, 3-problem round, one 2025 SRI-N drive used 50 MCQs, and one 2023 SRI-N intern drive shortlisted on resumes alone.

Do I need dynamic programming?

Some. DP appeared in 5 of 31 reports, but Burst Balloons was the whole 3-hour test in two of them. Graph and grid problems were more common (16 of 31).

Is system design asked?

Not as a separate round in any report. Design came up inside technical interviews: an LRU cache (2 reports), an LFU cache, a URL shortener at concept level, and diagrams of your own project.

How long does it take?

On campus, it can be a single day. One intern heard about the shortlist about an hour after the test, and one full-time drive held interviews the same afternoon. No report gives an end-to-end timeline in weeks.

Will the interviewer help me?

Sometimes. One research intern got a hint toward the dataset when stuck on overfitting, and another was told their grid approach was close to optimal. One interviewer also cut a candidate off for talking about development skills instead of AI.

Do interns get a pre-placement offer?

It depends on the track. SSIR described performance-based conversion after a 6-month internship, but an interviewer at a 2026 research-intern drive said Samsung usually does not offer interns a PPO.

About this guide

The Codinzhub team researched and wrote this guide from 32 Samsung software engineer interview experiences, mostly from 2023 to 2026, including experiences and coding-test questions shared on Codinzhub. We kept only what candidates described. The questions are our own short summaries, and the practice links are similar problems, not the exact interview questions unless a row says so. If you interviewed recently, share your experience on Codinzhub and we will update this guide.

Disclaimer. Samsung's process changes by centre, track, year and campus. Use this guide to build a foundation, and confirm the current format with your placement cell or recruiter.

Search the question bank

The questions from each round above, with practice links for similar problems. Log in to search every question we collected for this role, including questions shared on Codinzhub.

Showing 20 of 103 questions

  • Spaceship Coin Collector

    A Samsung online assessment question shared on Codinzhub (Samsung R&D, college not shown).

    On CodinzhubSamsung R&D, college not shownNot reported (grid problem)CodingOnline assessment
  • Doctor's Location Probability

    A Samsung online assessment question shared on Codinzhub (Samsung R&D, college not shown).

    On CodinzhubSamsung R&D, college not shownGraphs (directed graph)CodingOnline assessment
  • Research Center Location

    A Samsung online assessment question shared on Codinzhub (Samsung R&D, college not shown).

    On CodinzhubSamsung R&D, college not shownNot reportedCodingOnline assessment
  • N machines with unique machine IDs

    A Samsung online assessment question shared on Codinzhub (Samsung R&D Noida, IIT (ISM) Dhanbad).

    On CodinzhubSamsung R&D Noida, IIT (ISM) DhanbadNot reportedCodingOnline assessment
  • 13

    A Samsung online assessment question shared on Codinzhub.

    On CodinzhubSpaceship Coin Collector (grid): Codinzhub OA page for Samsung R&D; the public description shows a grid whose cells may hold different items.CodingOnline assessment
  • 14

    A Samsung online assessment question shared on Codinzhub.

    On CodinzhubDoctor's Location Probability on a directed graph: Codinzhub OA page for Samsung R&D: a doctor moves between divisions connected as a directed graph.CodingOnline assessment
  • 15

    A Samsung online assessment question shared on Codinzhub.

    On CodinzhubResearch Center Location: place a centre to minimise research cost: Codinzhub OA page for Samsung R&D: rare elements are found in a region and the team wants to minimise overall...CodingOnline assessment
  • 16

    A Samsung online assessment question shared on Codinzhub.

    On CodinzhubN machines with unique machine IDs (statement truncated publicly): Codinzhub OA page from a Samsung R&D Noida drive: you are given N machines, each with a unique machine ID.CodingOnline assessment
  • Burst Balloons: maximise coins by choosing the burst order

    Level: New grad.

    Interval DPCodingRound not stated
  • Find n where a monotonic equation f(n) equals a given value

    Level: SWE.

    Binary search over a monotonic functionCodingRound not stated
  • Graph problem built around cycle detection

    Level: Intern.

    Cycle detection in a graphCodingRound not stated
  • Flip whole columns exactly k times to maximise all-ones rows

    Level: Intern.

    Grouping identical rows, parity of leftover flipsCodingRound not stated
  • Kim and his Refrigerators (problem named; statement not given in the report)

    Level: New grad.

    Not reportedCodingRound not stated
  • Count substrings of s that read as a power of 2 (pen-and-paper round)

    Level: SWE.

    Not reportedCodingRound not stated
  • Grid DP with backtracking, similar to Cherry Pickup II

    Level: Intern.

    Multi-agent grid DPCodingRound not stated
  • Subtract two numbers stored as linked lists

    Level: Intern.

    Digit-by-digit arithmetic on listsCodingRound not stated
  • Empty cell farthest from its nearest city on a grid

    Level: Intern.

    Asked open-ended first: where to put a lab so the cities stay safeCodingRound not stated
  • Word Ladder: fewest one-letter changes from source to target

    Level: Intern.

    Not reportedCodingRound not stated
  • Decode a nested k[encoded] string

    Level: Intern.

    Not reportedCodingRound not stated
  • Longest consecutive sequence in O(n)

    Level: Intern.

    Not reportedCodingRound not stated
Page 1 of 6

Keep preparing on Codinzhub

Ready for the next topic?

Keep track of what you’ve learned.