Interview experience
Unacademy Interview Questions and Answers (2026)
Overview
Section titled “Overview”Unacademy’s fresher loop is a 4-5 stage process where coding rounds stay close to standard DSA, but a dedicated project deep-dive and a discussion-style managerial round carry real weight alongside the technical bar.
Unacademy interview process at a glance
Section titled “Unacademy interview process at a glance”| Round | Duration | What it tests |
|---|---|---|
| Online Assessment | ~90 min | 2 coding problems (strings/arrays, e.g. palindrome or frequency-counting patterns) |
| Technical Interview(s) | 45 min each | DSA (incl. graphs), plus a detailed project deep-dive |
| System design / SQL round (some loops) | 45 min | Lightweight system design, SQL, data-structure fundamentals |
| Managerial round | 45 min | Discussion-style behavioral - ownership, ambiguity |
| HR round | 30 min | Offer, logistics, general fit |
Online Assessment
Section titled “Online Assessment”Around 90 minutes with 2 coding problems, string/array-heavy. Reported prompts include longest palindromic substring and counting the occurrences of every element in a list - pattern-recognition problems more than exotic ones.
Common questions
- Longest palindromic substring
- Count occurrences of every element in a given list
- Basic array/string manipulation with edge-case follow-ups
Technical Interview(s)
Section titled “Technical Interview(s)”A DSA round (including graph problems) paired with - or followed by - a dedicated project-discussion segment that can run up to 45 minutes on its own. Some loops add a further round on system design, SQL, and core data structures.
Common questions
- Graph traversal problems (BFS/DFS-style)
- Detailed walkthrough of a resume project - architecture, stack choice, hardest bug
- Basic system-design prompt scoped to a bounded feature (some loops)
- SQL query questions and binary tree/hashmap/heap fundamentals (some loops)
Round-by-round narratives are on the Unacademy interview experience page.
Managerial round
Section titled “Managerial round”Described by candidates as discussion-oriented rather than a rapid Q&A session - the interviewer wants to see how you think and communicate, not just tick boxes on behavioral prompts.
Common questions
- How would you react if a project you’d worked on for months got shelved overnight
- Tell me about a time you had to learn something new quickly to unblock yourself or your team
- Describe how you’d prioritize when requirements shift mid-sprint
- Walk me through a disagreement with a teammate and how it resolved
HR round
Section titled “HR round”A closing conversation on logistics, offer details, and general fit once the technical and managerial bars are cleared.
Common questions
- Tell me about yourself and why Unacademy
- What do you know about Unacademy’s business beyond the app
- Are you comfortable with the pace of change in a company that has restructured recently
- Any questions about the team or role
Sample answer frameworks for each of these are on the Unacademy HR interview questions page.
Unacademy’s creator-economy model and recent restructuring
Section titled “Unacademy’s creator-economy model and recent restructuring”Unlike a pure SaaS or e-commerce company, Unacademy’s core product depends on a marketplace of educators (creators) producing live and recorded content - so domain questions sometimes probe how you’d think about content delivery at scale, live-class infrastructure, or learner engagement, rather than a generic CRUD system. It’s also worth knowing the company cut roughly 2,000 roles across multiple rounds of layoffs between late 2022 and mid-2024 (including 250 in mid-2024) as post-pandemic online-learning demand cooled and the company pushed toward profitability - the interview process itself hasn’t changed, but it’s reasonable to ask about team stability directly.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: How do you find the longest palindromic substring efficiently?
The standard interview solution is expand-around-centre: for each of the 2n-1 possible centres (n single characters and n-1 gaps between characters), expand outwards while the left and right characters match, and track the longest span seen. That runs in O(n^2) time and O(1) extra space, which is what Unacademy interviewers usually accept. The dynamic-programming table version is also O(n^2) time but needs O(n^2) space, so it is strictly worse. If asked for optimal, mention Manacher’s algorithm, which transforms the string with separators and reuses previously computed palindrome radii to get O(n) time.
Q: How would you count the occurrences of every element in a list?
Use a hash map from element to count: iterate the list once, and for each item increment its entry (defaulting to zero). That is O(n) time and O(k) space where k is the number of distinct values, and it beats the naive nested-loop count which is O(n^2). In Python this is collections.Counter(lst); in Java it is a HashMap with merge(x, 1, Integer::sum). If the values are small bounded integers, an array of size max+1 used as a direct-index counter is faster still because it avoids hashing entirely.
Q: What is the difference between BFS and DFS, and when do you pick each?
BFS explores level by level using a queue, so on an unweighted graph the first time it reaches a node it has found the shortest path in number of edges. DFS goes as deep as possible using a stack or recursion, which makes it natural for cycle detection, topological sort, and connected-component labelling. Both are O(V+E) time on an adjacency list; BFS space is proportional to the width of the frontier while DFS space is proportional to the recursion depth. For a shortest-hop question pick BFS, for a reachability or ordering question pick DFS.
Q: How do you detect a cycle in a directed graph?
Run DFS and keep three states per node: unvisited, in-recursion-stack, and fully processed. If DFS reaches a node that is currently in the recursion stack, you have found a back edge and therefore a cycle. This is O(V+E) time and O(V) space. The alternative is Kahn’s algorithm for topological sort: repeatedly remove nodes with in-degree zero, and if fewer than V nodes come out, the leftover nodes form a cycle. Note that the simple visited-set trick used for undirected graphs is not sufficient here, because a cross edge to an already-visited node is not a cycle in a directed graph.
Q: How does a hash map achieve O(1) average lookup, and when does it degrade?
A hash map applies a hash function to the key to compute a bucket index, so the expected number of keys per bucket stays small as long as the load factor is kept low (Java’s HashMap resizes at 0.75). Average get and put are therefore O(1). Degradation happens when many keys collide into one bucket - with a poor hash or adversarial keys, lookup becomes O(n) in the worst case for a plain chained implementation. Java 8 and later mitigates this by converting a bucket’s linked list into a balanced tree once it exceeds eight entries, bounding the worst case at O(log n).
Q: Write a SQL query to find the top 3 courses by number of enrolments.
Assuming an enrolments table with a course_id column, the query is SELECT course_id, COUNT(*) AS total FROM enrolments GROUP BY course_id ORDER BY total DESC LIMIT 3;. The GROUP BY collapses rows per course, COUNT() gives the enrolment count, and ORDER BY with LIMIT takes the top three. Note you cannot put the aggregate condition in WHERE - filters on COUNT() must go in HAVING, because WHERE is evaluated before grouping. On SQL Server the equivalent is SELECT TOP 3 instead of LIMIT.
Q: What is the difference between a min-heap and a balanced BST for finding the smallest element?
A binary min-heap gives you the minimum in O(1) by reading the root, with O(log n) insert and O(log n) extract-min, but it has no useful ordering beyond the root - searching for an arbitrary value is O(n). A balanced BST gives O(log n) search, insert, delete, and in-order traversal in sorted order, but reading the minimum still costs O(log n) as you walk left. So for a top-k or streaming-minimum problem use a heap; for range queries or ordered iteration use a BST. Heaps also have better constants because they live in a contiguous array with no pointer chasing.
Q: How do you decide between SQL and NoSQL when designing a feature like live-class attendance?
Attendance writes arrive in a heavy, append-only burst when a live class starts, and each record is small and independent - that suits a wide-column or document store that partitions by class ID and scales writes horizontally. A relational database is the better choice when you need multi-row transactions and joins, for example reconciling subscriptions against payments where correctness matters more than write throughput. In practice you often use both: the relational store as the source of truth for billing and enrolment, and a NoSQL store or event log for high-volume telemetry. Say the trade-off out loud in the round - interviewers are grading the reasoning, not the label.
Frequently asked questions about Unacademy interviews
Section titled “Frequently asked questions about Unacademy interviews”What is the Unacademy interview process for freshers?
Unacademy’s SDE loop typically runs 4-5 stages: 1. Online Assessment (~90 min) - 2 coding problems (candidates report classics like longest palindromic substring and counting occurrences of each element). 2. Technical Interview - DSA problems including graph questions, plus a detailed project-discussion round. 3. A further technical round some candidates report as System Design, Algorithms, Data Structures, and SQL combined - up to 5 coding problems across the loop. 4. Managerial round - primarily behavioral, discussion-style rather than Q&A. 5. HR round - offer and logistics. Total timeline is roughly 3-4 weeks.
What questions are asked in Unacademy interviews?
Coding rounds lean on strings, arrays, and graphs - longest palindromic substring, counting element occurrences, and graph traversal problems are commonly reported. A separate round goes deep on a resume project (up to 45 minutes). Some loops add SQL and binary tree/hashmap/heap questions. The managerial round is explicitly discussion-oriented rather than rapid-fire Q&A, and covers ownership and how you handle ambiguity in a fast-changing edtech product.
How many rounds are there in the Unacademy interview?
Most reports describe 4-5 touchpoints: an online assessment, one or two DSA/project-focused technical rounds, sometimes a dedicated system-design-plus-SQL round, a managerial/behavioral round, and a closing HR round. The exact count varies by team and hiring cycle - some drives compress stages.
How should I prepare for Unacademy interviews?
Practice string and array problems (palindrome and frequency-counting patterns come up specifically) plus graph traversal for the technical rounds. Prepare one project you can discuss for a full 45 minutes - stack choices, trade-offs, what broke. Revise basic SQL and be ready for a lightweight system-design prompt. For the managerial round, prepare stories on working through a deprioritized project or a fast pivot, since Unacademy’s content/creator business has gone through real strategy shifts.
Has Unacademy’s edtech slowdown affected its hiring or interviews?
Unacademy cut roughly 2,000 roles across several rounds of layoffs between late 2022 and mid-2024 as post-pandemic demand for online learning cooled and the company pushed toward profitability, including a 250-person cut in mid-2024. This hasn’t changed the interview format, but it’s a reasonable thing to ask about directly - which team you’d join and how stable it currently is - rather than assuming steady growth-stage hiring.
What does Unacademy’s technical round test beyond DSA?
Beyond coding, expect a genuine project deep-dive (not a quick summary) and, on some loops, a lightweight system-design or SQL segment. Interviewers weigh how clearly you explain trade-offs and handle follow-up probing over whether you memorized a pattern.