Interview experience
Samsung Interview Questions and Answers (2026)
Overview
Section titled “Overview”Samsung R&D Institute India runs large campus and off-campus drives with a strict, hidden-test-case coding assessment followed by CS-fundamentals technical rounds - not a Korea-style GSAT process.
Samsung interview process at a glance
Section titled “Samsung interview process at a glance”| Round | Duration | What it tests |
|---|---|---|
| Online Coding Assessment | 2-3 hrs | 1-3 DSA problems, all/most hidden test cases must pass; sometimes + technical MCQs |
| Technical Interview (1-2 rounds) | 45-60 min each | DSA (explained verbally), OS, DBMS, OOP, networking, C/C++, projects |
| HR Round | 20-30 min | Fit, motivation, occasionally a puzzle; usually lighter than the technical rounds |
Online Coding Assessment
Section titled “Online Coding Assessment”The first filter is a 2-3 hour online round, sometimes run on Samsung’s own platform with an external proctoring vendor. Expect 1-3 DSA problems (graph problems, especially cycle detection, come up often) with a hidden test suite that is stricter than most fresher assessments - a near-complete solution that fails even one hidden case typically doesn’t advance. Some cycles restrict STL/Collections usage; others don’t, so confirm the rules on the day rather than assuming.
Common questions
- Detect a cycle in a directed graph (and identify the nodes forming it)
- Array/string problems solved under a strict time and test-case bar
- A graph-shortest-path or traversal problem in a real-world framing (e.g. “wormhole”-style problems have shown up)
- Technical MCQs on OS, DBMS, and OOP (on drives that add this section)
Full round-by-round breakdowns are on the Samsung interview experience page.
Technical Interview
Section titled “Technical Interview”Candidates who clear the coding assessment usually face one or two technical interviews. DSA is often discussed verbally - explain your approach and complexity rather than writing full code - alongside core CS fundamentals: operating systems, DBMS, OOP concepts, computer networks, and C/C++ specifics. Interviewers also spend real time on your resume projects, so be ready to defend design choices, not just describe what you built.
Common questions
- Explain your approach to a medium-hard DSA problem (linked list, tree, or stack based) without writing code
- Process vs thread, deadlock conditions, and paging/segmentation (OS)
- Normalization, indexing, and joins (DBMS)
- OOP pillars with a concrete example from your own project
- TCP vs UDP, and what happens when you type a URL in a browser (networking)
- Walk through a design decision in your final-year or internship project
See how real candidates handled this stage on the Samsung interview experience page.
HR Round
Section titled “HR Round”The HR round at Samsung R&D India tends to be shorter and less adversarial than the technical rounds, sometimes done in small groups. It’s still an evaluation, not a pure formality - vague or rehearsed answers can cost you even here.
Common questions
- Tell me about yourself
- Why Samsung, and why R&D specifically?
- Walk me through a project where you had to optimize a solution to pass every test case
- Tell me about a time you worked in a fast-changing technical environment
- A short logic puzzle, on some drives
Sample answer frameworks for each of these are on the Samsung HR interview questions page.
GSAT vs the SRI engineering pipeline
Section titled “GSAT vs the SRI engineering pipeline”Samsung’s Global Samsung Aptitude Test (GSAT) is a real thing, but it’s easy to conflate with the SRI hiring process. GSAT is primarily used for Samsung’s broader group-level and business/sales hiring (and heavily for hiring in Korea), covering quantitative ability, reasoning, and general aptitude in a timed format. Most software engineering candidates applying to Samsung R&D Institute India go through the coding-assessment-plus-technical-interview pipeline described above instead - if a job posting or recruiter mentions GSAT for an SRI software role, confirm directly rather than assuming it replaces the coding round.
Samsung PRISM: a common feeder route
Section titled “Samsung PRISM: a common feeder route”Samsung PRISM pairs students (typically final-year or pre-final-year, CGPA 7+) with Samsung engineers on real industry-academia projects, and strong performers are a recognized source of full-time offers and internships at SRI Bangalore. PRISM’s own selection process is usually three stages: a coding round (arrays, trees, linked lists), a basic aptitude test, and an interview focused on personality and motivation rather than deep technical depth.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: How do you detect a cycle in a directed graph and identify the nodes forming it?
Run DFS keeping two arrays: visited, and inRecursionStack. When DFS reaches a node already marked inRecursionStack, that edge is a back edge and a cycle exists. To recover the cycle itself, keep a parent array and walk back from the current node through parents until you return to the node you hit, collecting the nodes along the way. Clear the inRecursionStack flag as each call returns, otherwise you will report false cycles across separate branches. The whole traversal is O(V + E) time and O(V) space. Kahn’s topological sort is the alternative: if it emits fewer than V nodes, the leftovers are exactly the nodes on cycles.
Q: How do you find the shortest path in a graph that has negative edge weights?
Dijkstra’s algorithm breaks with negative edges because it finalises a node’s distance the moment it is popped, which a later negative edge can invalidate. Use Bellman-Ford instead: initialise the source distance to 0 and all others to infinity, then relax every edge V-1 times, since any shortest path has at most V-1 edges. Run a V-th pass; if any distance still improves, the graph has a negative-weight cycle and no shortest path exists. Complexity is O(V times E) time and O(V) space. For all-pairs shortest paths with negative edges, Floyd-Warshall gives O(V cubed) and flags a negative cycle by a negative value on the diagonal.
Q: What is the difference between paging and segmentation?
Paging divides both logical and physical memory into fixed-size frames and pages, typically 4 KB, so a logical address splits into a page number and an offset that indexes a page table. It eliminates external fragmentation completely, but causes internal fragmentation in the last partial page, and it is invisible to the programmer. Segmentation divides memory into variable-length, logically meaningful units - code, stack, heap - so an address is a segment number plus an offset checked against that segment’s limit. It matches the program’s structure and simplifies protection and sharing per segment, but produces external fragmentation. Modern x86 systems use segmentation with paging, where each segment is itself paged.
Q: What happens when you type a URL into a browser and press Enter?
The browser parses the URL and checks its cache, then resolves the hostname: OS cache, then the configured resolver, which walks root, TLD, and authoritative nameservers if needed. With an IP in hand it opens a TCP connection through the three-way handshake, and for HTTPS performs a TLS handshake that negotiates a cipher, validates the server certificate chain, and derives session keys. It then sends an HTTP request; the server may respond with a redirect or a 200 with HTML. The browser parses HTML into the DOM and CSS into the CSSOM, blocking on synchronous scripts, then combines them into a render tree, runs layout and paint, and fetches sub-resources - often over the same connection thanks to keep-alive or HTTP/2 multiplexing.
Q: Why do database indexes use B+ trees rather than binary search trees?
Disk and SSD reads happen in pages, so the cost that matters is the number of page reads, not comparisons. A B+ tree keeps hundreds of keys per node, so its height stays around three or four even for hundreds of millions of rows, meaning a lookup costs only a few I/Os, while a balanced binary tree of the same size would be roughly 27 levels deep. B+ trees also store all data pointers in the leaves and chain the leaves in a linked list, which makes range scans and ORDER BY sequential rather than random. A clustered index stores the rows themselves in that leaf order, so a table has only one; non-clustered indexes store a pointer back to the row and may need an extra lookup unless the index covers the query.
Q: What is the difference between malloc/free and new/delete in C++?
malloc allocates a raw block of bytes and returns void star, so you must cast it, compute the size yourself, and check for a NULL return; free simply releases the block. new computes the size from the type, returns a correctly typed pointer, invokes the constructor, and throws std::bad_alloc on failure rather than returning NULL; delete invokes the destructor before releasing memory. Mixing them is undefined behaviour, as is using delete on memory from new[] - that needs delete[] so every element’s destructor runs. In modern C++ you should be reaching for std::unique_ptr, std::shared_ptr, and std::vector rather than any of these directly.
Q: How do you reverse a linked list in groups of k nodes?
Process the list one block at a time. First check that at least k nodes remain by walking k steps ahead; if not, leave the tail as is (or reverse it too, depending on the problem statement). Reverse those k nodes with the standard three-pointer iterative reversal, which leaves the block’s original head as its new tail. Connect the previous block’s tail to the new head, then recurse or loop onto the node that followed the block. Each node is visited a constant number of times, so it is O(n) time; the iterative version is O(1) space while the recursive one costs O(n/k) stack frames. Using a dummy head node avoids special-casing the first block.
Q: How would you design a stack that returns its minimum element in O(1)?
Keep a second stack of minimums alongside the main stack. On push, also push the smaller of the new value and the current minimum onto the min stack; on pop, pop both. getMin then reads the top of the min stack in O(1), at the cost of O(n) extra space. The space-optimised variant uses a single stack storing an encoded value: when a new minimum arrives, push two times the value minus the current minimum, update the minimum, and decode on pop by restoring the old minimum as two times the current minimum minus the stored value. That gets O(1) space overhead but risks integer overflow, which is worth stating explicitly.
Frequently asked questions about Samsung interviews
Section titled “Frequently asked questions about Samsung interviews”What is the Samsung R&D India interview process for freshers?
Samsung R&D Institute India (SRI-Bangalore, SRI-Noida, SRI-Delhi) typically runs 3-4 rounds: 1. Online Coding Assessment (2-3 hours) - 1-3 DSA problems where hidden test cases must pass, sometimes with technical MCQs added. 2. Technical Interview (one or two rounds, 45-60 min each) - DSA discussed verbally plus OS, DBMS, OOP, networking, and C/C++ fundamentals. 3. HR Round - fit questions and occasionally a puzzle or two, generally lighter than the technical rounds. Total timeline is usually 2-4 weeks for on-campus drives.
What questions are asked in Samsung interviews?
Coding assessments lean on arrays, strings, linked lists, trees, and graph problems (cycle detection is a recurring theme), graded against a full hidden test suite rather than the visible samples. Technical interviews add core CS fundamentals - OS, DBMS, OOP, computer networks - and a walkthrough of your resume projects. Some cycles restrict STL/Collections usage, though this varies by campus and year.
How many rounds are there in the Samsung interview?
Most SRI drives run 3-4 rounds for freshers: an Online Coding Assessment, one or two Technical Interviews, and an HR round. Some cycles report an extra screening round, taking the total to 5. Exact structure varies by location (Bangalore, Noida, Delhi) and hiring season.
What is GSAT and does it apply to Samsung R&D India hiring?
GSAT (Global Samsung Aptitude Test) is Samsung’s broader group-level aptitude test, used mainly for Samsung’s business/sales/corporate hiring tracks and for hiring in Korea - it is a separate pipeline from Samsung R&D Institute India’s engineering recruitment. Most SRI software engineering candidates in India go through the coding-and-technical-interview process described on this page, not GSAT, though the two can be confused since both fall under the Samsung umbrella.
What is Samsung PRISM?
Samsung PRISM is an academia-industry research internship program that pairs students with Samsung engineers on real projects, and is one of the more common feeder routes into a full-time offer at SRI Bangalore. Its own selection process is typically three stages: a coding round (arrays, trees, linked lists), a basic aptitude test, and a personality/motivation-focused interview.
How should I prepare for Samsung interviews?
Practice DSA problems until they pass every hidden test case (not just the visible samples) under a strict time limit, revise OS/DBMS/OOP/networking fundamentals since they come up in nearly every technical round, and prepare a clear, honest walkthrough of your resume projects. Since the HR round is comparatively light, weight your prep toward the coding assessment and technical interviews.