Interview experience
Adobe Interview Questions and Answers (2026)
Overview
Section titled “Overview”Adobe runs real campus and off-campus drives out of Noida and Bangalore, pairing DSA-heavy technical rounds with a values-driven Hiring Manager or Director conversation.
Adobe interview process at a glance
Section titled “Adobe interview process at a glance”| Round | Duration | What it tests |
|---|---|---|
| Online Assessment | 60-90 min | Aptitude/logical reasoning + 2-4 DSA coding problems + CS-fundamentals MCQs |
| Technical Interview 1 | 45-60 min | Live DSA coding on a shared editor, complexity trade-offs |
| Technical Interview 2 | 45-60 min | DSA + CS fundamentals (OS/DBMS/networking) + project deep-dive |
| Hiring Manager / Director round | 45-60 min | Project ownership, technical judgment, values-driven behavioural questions |
| HR round | 20-30 min | Fit, logistics, offer discussion |
Online Assessment
Section titled “Online Assessment”A HackerRank-hosted test that runs 60-90 minutes: an aptitude/logical-reasoning section (seating arrangements, data interpretation, quantitative aptitude) followed by 2-4 DSA coding problems and a batch of C/C++/Java or CS-fundamentals MCQs. Most drives expect all visible test cases to pass on the coding problems, not just partial credit.
Common questions
- Longest palindromic substring / minimum window substring
- LRU or LFU cache implementation
- Maximum height of a tower buildable between two given towers (greedy/math)
- OOP, OS, and DBMS MCQs alongside the coding problems
Full round-by-round breakdowns are on the Adobe interview experience page.
Technical Interview 1
Section titled “Technical Interview 1”A 45-60 minute live coding round, usually over video call with a shared editor. Expect 1-2 DSA problems with follow-up questions on optimizing space/time complexity and handling edge cases - Adobe interviewers generally want to hear your reasoning as you go, not just a final answer.
Common questions
- Serialize and deserialize a binary tree
- Trapping rain water (two-pointer)
- Word ladder / shortest transformation sequence (BFS)
- Course schedule I/II (topological sort)
See how real candidates handled this stage on the Adobe interview experience page.
Technical Interview 2
Section titled “Technical Interview 2”Another 45-60 minute round, but weighted more toward CS fundamentals and your resume project alongside a DSA problem. Some loops add a lightweight system-design discussion here - designing a document-storage or notification system in plain language, not a full distributed-systems deep-dive.
Common questions
- Process vs thread, virtual memory, paging
- Database indexing and B-trees
- Design basics for a document-storage or notification system
- Deep dive into your internship/major project: design choices and trade-offs
Hiring Manager / Director round
Section titled “Hiring Manager / Director round”Adobe’s most distinctive technical-adjacent round: a 45-60 minute conversation, often with a Director or senior engineering manager, that mixes a project deep-dive with behavioural questions tied to Adobe’s culture. Interviewers probe the “why” behind your technical decisions - not just what you built - and look for ownership of outcomes, including ones that didn’t go as planned.
Common questions
- Walk me through a design decision you’d change today, and why
- Tell me about a time you had to defend a technical choice to someone senior
- How do you handle disagreements within a team?
- Where do you see yourself in five years?
Sample answer frameworks for each of these are on the Adobe HR interview questions page.
HR round
Section titled “HR round”A short 20-30 minute closing conversation covering fit, location/relocation, notice period, and offer logistics. On some campus drives this gets folded directly into the Hiring Manager/Director round rather than run separately.
Common questions
- Tell me about yourself
- Why Adobe?
- Are you willing to relocate to Noida/Bangalore?
- Do you have any questions for us?
Adobe hires at real scale in India
Section titled “Adobe hires at real scale in India”Adobe runs genuine large-scale on-campus and off-campus drives out of Noida and Bangalore, typically gated by a 7.0+ CGPA cutoff for CSE/IT/ECE branches - a real placement-season pipeline, not a referral-only trickle. Because its business units (Creative Cloud, Document Cloud, Experience Cloud) each run their own variant of the loop, round count and exact question mix can differ even within the same overall drive.
Adobe’s “Check-in” culture
Section titled “Adobe’s “Check-in” culture”Adobe is well known in HR circles for scrapping annual performance reviews in 2012 in favor of frequent, informal “Check-in” conversations between managers and employees - no forced stack ranking, no once-a-year scramble. This isn’t a formal interview round, but it’s a genuine piece of Adobe’s culture that experienced interviewers sometimes reference when asking how you like to receive feedback - candidates who frame their answer around continuous, informal feedback (rather than annual review cycles) tend to land better than those who don’t know the reference.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: How do you design an LRU cache with O(1) get and put?
Combine a hash map with a doubly linked list. The hash map maps each key to the address of its node, giving O(1) lookup; the doubly linked list keeps nodes in recency order with the most recently used at the head and the least recently used at the tail. On get, look the key up in the map, unlink its node and re-insert it at the head, and return the value. On put, if the key exists update the value and move the node to the head; otherwise create a node at the head, and if the size now exceeds capacity remove the tail node and delete its key from the map. A doubly linked list is essential because unlinking a node given only a pointer to it needs both neighbours, which is O(1) here and O(n) with a singly linked list. Both operations are O(1) time and the structure is O(capacity) space. LFU differs by keying on frequency: you keep a map from frequency to a list of keys plus a minFreq pointer to still get O(1).
Q: Explain the two-pointer solution to Trapping Rain Water.
Water above index i equals min(maxLeft[i], maxRight[i]) - height[i], where maxLeft and maxRight are the tallest bars to the left and right. The naive approach precomputes both arrays in O(n) time and O(n) space. The two-pointer version removes the arrays: keep pointers left at 0 and right at n-1 plus running leftMax and rightMax. At each step, if height[left] < height[right] then leftMax is guaranteed to be the binding constraint for the left pointer, because some bar at least as tall as height[right] exists on the right; so add leftMax - height[left] to the answer and advance left, otherwise do the symmetric thing on the right. Each pointer moves at most n steps, so the algorithm is O(n) time and O(1) extra space. Adobe interviewers usually ask you to justify why the smaller side is safe to settle first.
Q: How do you find the longest palindromic substring efficiently?
The expand-around-centre approach treats every one of the 2n-1 possible centres, n single characters and n-1 gaps between characters, and expands outward while the characters on both sides match, tracking the longest span found. That is O(n squared) time and O(1) space, and it is the answer most interviewers expect. The dynamic-programming formulation, where dp[i][j] is true when s[i] == s[j] and dp[i+1][j-1] is true, is also O(n squared) time but costs O(n squared) space, so it is strictly worse in practice. If pushed for optimal, Manacher’s algorithm achieves O(n) by inserting separator characters to make every palindrome odd-length and reusing previously computed radii mirrored about the current centre, so each position is expanded past only once.
Q: How does topological sort solve Course Schedule II, and how do you detect a cycle?
Model courses as vertices and each prerequisite pair as a directed edge from prerequisite to course, then output any topological order. Kahn’s algorithm computes the in-degree of every vertex, pushes all zero in-degree vertices onto a queue, and repeatedly pops a vertex, appends it to the result, and decrements the in-degree of each neighbour, enqueueing any that hit zero. If the result contains fewer than V vertices when the queue empties, the remaining vertices sit on a cycle and no valid ordering exists, which is exactly the Course Schedule I answer. The DFS alternative colours vertices white, grey, and black; encountering a grey vertex means a back edge and therefore a cycle, and pushing vertices onto a stack at post-order gives the reverse topological order. Both run in O(V + E) time and O(V + E) space.
Q: Why do relational databases use B+ trees for indexes rather than hash indexes or binary search trees?
A B+ tree is a balanced multi-way tree where internal nodes hold only keys for routing and all actual records or row pointers live in the leaves, which are chained in a linked list. Node size is matched to the disk page, typically 4 to 16 KB, so a node holds hundreds of keys and the tree stays about 3 to 4 levels deep even for hundreds of millions of rows, meaning a lookup costs only a handful of page reads. A binary search tree has fan-out 2, so its depth would be around 27 for 100 million rows and every level is a potential disk seek. A hash index gives O(1) equality lookups but cannot serve range scans or ORDER BY, whereas the B+ tree’s linked leaf level makes range queries a sequential walk. Adobe follow-ups usually cover clustered versus secondary indexes and why a leading-column mismatch stops a composite index from being used.
Q: Explain virtual memory, paging, and what happens on a page fault.
Virtual memory gives each process its own contiguous address space that the MMU translates to physical frames through per-process page tables, providing isolation and allowing the total working set to exceed RAM. Memory is split into fixed-size pages, typically 4 KB, mapped to equal-size physical frames, which eliminates external fragmentation at the cost of some internal fragmentation in the last page. Because walking a multi-level page table on every access would be slow, the CPU caches recent translations in the TLB. A page fault is the trap raised when the page table entry is marked not-present: the OS checks whether the access is legal, and if so finds a free frame or evicts one using an approximation of LRU such as the clock algorithm, writes the victim back if its dirty bit is set, loads the page from disk, updates the page table, and restarts the faulting instruction. Sustained faulting where processes spend more time swapping than executing is called thrashing.
Q: What is the difference between HTTP and HTTPS, and what happens during a TLS handshake?
HTTPS is HTTP carried inside a TLS session, so it adds encryption to prevent eavesdropping, integrity checking to prevent tampering, and server authentication through an X.509 certificate chain that terminates in a trusted certificate authority. In a TLS 1.3 handshake the client sends a ClientHello with supported cipher suites and a key-share for a guessed group, the server replies with its ServerHello, key-share, certificate, and a CertificateVerify signature, both sides derive the shared secret via ephemeral Diffie-Hellman on an elliptic curve, and application data flows after one round trip rather than TLS 1.2’s two. Ephemeral keys give forward secrecy, so recording the traffic and later stealing the server’s private key does not decrypt it. HTTPS is also a prerequisite for HTTP/2 and HTTP/3 in every mainstream browser, so it usually improves performance rather than costing it.
Q: How would you design a notification system that delivers push, email, and SMS?
Accept notification requests at an API service that validates and de-duplicates using an idempotency key, then write each request to a durable queue such as Kafka partitioned by user ID so a single user’s notifications keep their ordering. A dispatcher service reads the queue, looks up user preferences and device tokens, applies rate limits and quiet hours, and fans out per-channel messages to separate worker pools for APNs and FCM push, an email provider, and an SMS gateway. Each worker retries with exponential backoff and jitter, and messages that exhaust retries land in a dead-letter queue for inspection. Store delivery state per notification so you can report sent, delivered, and opened, and use a template service so message bodies are versioned rather than hard-coded. The main trade-offs an Adobe interviewer will probe are at-least-once delivery plus consumer-side idempotency versus exactly-once, and how you keep a burst of millions of marketing notifications from starving latency-sensitive transactional mail.
Frequently asked questions about Adobe interviews
Section titled “Frequently asked questions about Adobe interviews”What is the Adobe interview process for SDE/MTS roles?
Adobe’s loop usually runs 4-6 stages: 1. An Online Assessment (60-90 min) - aptitude/logical reasoning plus 2-4 DSA coding problems and CS-fundamentals MCQs (OOP, OS, DBMS, networking), run on HackerRank. 2. Two technical interviews (45-60 min each) - live DSA coding on a shared editor, then a second round mixing DSA with project deep-dive and sometimes basic system design. 3. A Hiring Manager or Director round (45-60 min) - project ownership, technical trade-offs, and values-driven behavioural questions. 4. A short HR round (20-30 min) for fit and offer logistics. Campus timelines run 2-4 weeks; off-campus/lateral loops often take 3-6 weeks and add a round.
What questions are asked in Adobe interviews?
Coding rounds lean LeetCode medium-to-hard: arrays/strings (sliding window, two-pointer), trees and graphs (BFS/DFS, topological sort), dynamic programming, and caching structures like LRU/LFU. CS-fundamentals questions cover OOP concepts, OS topics (virtual memory, paging, process vs thread), DBMS (indexing, normalization), and networking (HTTP/HTTPS, DNS, load balancers). The Hiring Manager/Director round probes your resume project in depth - the “why” behind design choices, not just the what - alongside basic system-design questions like designing a document-storage or notification system.
How many rounds are there in the Adobe interview?
Typically 4-6 touchpoints: an Online Assessment, two technical/coding interviews, a Hiring Manager or Director round, and a closing HR round. Some campus drives compress this to 4 rounds by folding HR questions into the Director round; off-campus and experienced-hire loops sometimes add an extra technical round.
How should I prepare for Adobe interviews?
Drill medium-to-hard DSA (arrays, trees, graphs, DP, caching structures) until you can explain complexity trade-offs out loud, revise OS/DBMS/networking fundamentals since Adobe’s fundamentals questions carry real weight, be ready to walk through your major project’s design decisions in depth, and prepare 2-3 STAR stories around ownership and impact for the Hiring Manager/Director round.
Does Adobe hire through campus placements in India?
Yes. Unlike some FAANG peers that hire India engineers mostly off-campus, Adobe runs genuine large-scale on-campus and off-campus drives out of its Noida and Bangalore offices, typically with a 7.0+ CGPA cutoff for CSE/IT/ECE branches. Business units (Creative Cloud, Document Cloud, Experience Cloud) hire in parallel, so exact round count and question style can vary by team even within the same drive.
What is Adobe’s “Check-in” culture, and does it come up in interviews?
Adobe famously replaced annual performance reviews with frequent, informal “Check-in” conversations between managers and employees back in 2012 - no forced rankings, no once-a-year scramble. It’s not a formal interview topic, but interviewers (especially in the Hiring Manager/Director round) sometimes ask how you prefer to receive feedback or handle ongoing coaching, and framing your answer around continuous feedback rather than annual reviews signals you’ve done real homework on Adobe specifically.
What compensation does Adobe offer freshers/SDE-1 in India?
Reported packages for fresher SDE/MTS roles cluster in the roughly Rs.20-45 LPA range (base plus stock/bonus), varying widely by college tier, CGPA, and business unit - treat any single number from a forum post as one data point, not a guarantee.