Interview experience
KPMG Interview Questions and Answers (2026)
Overview
Section titled “Overview”KPMG Global Services (KGS, KPMG’s technology delivery arm and the entry point for most CS/IT campus hires) runs a 2-4 stage process built around a five-section timed aptitude test, a data-structures-and-project technical interview, and a closing HR round.
KPMG interview process at a glance
Section titled “KPMG interview process at a glance”| Round | Duration | What they test |
|---|---|---|
| Online Aptitude Test | 60 min | Math ability, logical reasoning, verbal ability, pseudocode, puzzles |
| Group Discussion (if conducted) | 20-30 min | Current affairs / business topic |
| Technical Interview | 30-45 min | Projects, data structures, algorithmic problem-solving, language trivia |
| HR Interview | 20-30 min | Consulting fit, 2-year plan, client-readiness |
Online Aptitude Test
Section titled “Online Aptitude Test”A strictly timed, five-section paper: Mathematics Ability, Logical Reasoning, Verbal Ability, Pseudocode, and Puzzles, each with its own time block and no navigating back once a section starts. No negative marking, but pacing within each section matters since you can’t return to skipped questions.
Common questions
- Mathematics: percentages, ratios, averages, time-and-work
- Logical reasoning puzzles and syllogisms
- Pseudocode: trace through algorithm-like code and predict output or spot the error
- Verbal ability: reading comprehension, grammar
- Classic logic puzzles (weighing, arrangement problems)
Group Discussion (if conducted)
Section titled “Group Discussion (if conducted)”Held at some campuses, 15-20 candidates typically split into groups of 6-8 to discuss a current-affairs or business topic for 15-20 minutes, evaluated on structured argument and listening, not volume.
Common questions
- Impact of AI on traditional consulting
- Digital transformation in Indian businesses
- A current business or technology news topic
Round-by-round breakdowns are on the KPMG interview experience page.
Technical Interview
Section titled “Technical Interview”Evaluates fundamental CS knowledge, data-structure proficiency, and algorithmic problem-solving alongside a resume/project discussion. Expect language-specific trivia and questions that test how you break down an ambiguous, open-ended problem.
Common questions
- Walk through your projects - stack, your specific contribution, hardest bug
- How would you approach a client’s data-migration project? (open-ended, tests structured thinking)
- Explain database normalization with an example
- Language-specific trivia - memory management, OOP paradigms, common gotchas
- Data-structure fundamentals - when to use a hash map vs a tree vs a queue
HR Interview
Section titled “HR Interview”Assesses consulting fit directly: a clear 2-year plan, one genuine growth area, and comfort with client-facing work, plus standard motivation and culture-fit questions.
Common questions
- Why KPMG, and why consulting?
- Tell me about a time you worked under tight deadlines, such as busy season - how did you manage it?
- How would you handle disagreeing with a senior colleague’s judgment on a client matter?
- What’s your 2-year plan, and where’s your biggest growth area right now?
- What do you know about KPMG’s values (Integrity, Excellence, Courage, Together, For Better)?
Sample answer frameworks for each of these are on the KPMG HR interview questions page.
Audit/Tax vs KGS/Technology: why the track matters
Section titled “Audit/Tax vs KGS/Technology: why the track matters”KPMG hires across genuinely different business lines - Audit & Assurance, Tax, Advisory/Consulting, and KGS (KPMG Global Services, the technology delivery arm running from eight India locations that does actual software, data, and analytics work). This page describes the KGS/technology track most CS/IT campus hires enter: an OA with a Pseudocode section, a data-structures-and-project technical interview, and HR. Audit and Tax hiring, aimed largely at commerce and accounting graduates, runs a lighter process built around aptitude, group discussion, and HR, with no coding component. Confirm which business line your offer letter or registration email names before you prep.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: In a pseudocode question, how do you work out the complexity of a nested loop?
Count how many times the innermost statement executes as a function of n, from the inside out. If the outer loop runs n times and the inner loop runs n times for each, that is n squared. But if the inner variable doubles each iteration - i starting at 1 and becoming i * 2 until it reaches n - the inner loop runs only log base 2 of n times, giving O(n log n) overall. If the inner loop’s bound depends on the outer index, as in an inner loop from 0 to i, the total is 1 plus 2 plus up to n, which is n times n plus one over two, and therefore still O(n squared). The trap KPMG’s Pseudocode section likes is a loop whose counter is modified inside the body, so always trace two or three iterations by hand with a small value such as n equals 4 before answering.
Q: When would you use a hash map versus a balanced tree versus a queue?
Use a hash map when you need average O(1) lookup, insert, and delete by exact key and you do not care about ordering - a lookup table of customer IDs, or counting occurrences. Use a balanced tree such as a TreeMap or a database B-tree index when you need the keys kept in sorted order, because it supports range queries, nearest-key lookups, and ordered iteration in O(log n), none of which a hash map can do at all. Use a queue when the requirement is processing order rather than lookup - FIFO for task processing, or a priority queue when the next item should be the highest-priority one, at O(log n) per insert and extract. The honest framing for an interviewer is that hash maps trade ordering for speed, trees trade a log factor for ordering, and queues are about scheduling rather than search.
Q: What is the difference between a primary key, a unique key, and a foreign key?
A primary key uniquely identifies each row, cannot be NULL, and there is exactly one per table; most engines create a clustered or primary index on it automatically. A unique key also enforces uniqueness but permits NULLs (typically one, depending on the engine) and a table can have several - an email column is the usual example, where you want no duplicates but the column is not the row’s identity. A foreign key is a column that references the primary key of another table, enforcing referential integrity so you cannot insert an order for a customer that does not exist, and controlling what happens on deletion via ON DELETE CASCADE, SET NULL, or RESTRICT. A composite key is simply a primary key made of more than one column, which is common in join tables.
Q: What is the difference between WHERE and HAVING in SQL?
WHERE filters individual rows before grouping and aggregation happen, so it cannot reference an aggregate function. HAVING filters the groups after GROUP BY has aggregated them, so it can. A query that needs both reads naturally: SELECT dept_id, COUNT(emp_id) FROM employees WHERE status = 'ACTIVE' GROUP BY dept_id HAVING COUNT(emp_id) > 5; - the WHERE removes inactive employees before counting, and the HAVING keeps only departments with more than five active employees. Doing the row filter in HAVING instead would be both wrong and slower, since every row would be aggregated first. The logical processing order worth memorising is FROM, WHERE, GROUP BY, HAVING, SELECT, ORDER BY, which also explains why a column alias defined in SELECT cannot be used in WHERE.
Q: What is the difference between stack and heap memory?
The stack holds function call frames - local variables, parameters, and return addresses - and is managed automatically: memory is allocated when a function is entered and freed when it returns, in strict LIFO order. It is very fast, since allocation is just moving a pointer, but it is small and fixed, which is why deep or infinite recursion produces a stack overflow. The heap is the region for dynamically allocated memory whose lifetime is not tied to a function call - new in Java and C++, malloc in C - so it is larger and flexible, but allocation is slower and it fragments over time. In C and C++ you must free heap memory yourself or you leak it; in Java and C# a garbage collector reclaims unreachable objects. The interview-ready summary: object references live on the stack, the objects themselves live on the heap.
Q: What are the four pillars of object-oriented programming?
Encapsulation bundles data with the methods that operate on it and restricts direct access to the fields, so an object controls its own invariants through getters and setters. Abstraction exposes only the essential interface and hides the implementation, so callers depend on what an object does rather than how. Inheritance lets a subclass reuse and specialise a parent’s behaviour, expressing an is-a relationship - though composition is usually preferred where the relationship is really has-a, since deep hierarchies become brittle. Polymorphism lets one interface serve many types, either at compile time through overloading or at run time through overriding, which is what allows new types to be added without changing the code that consumes them. Interviewers usually want a one-line example for each rather than the definitions alone.
Q: How would you approach a client’s data-migration project?
Start by profiling the source: row counts, distinct values, null rates, and the actual data types, since legacy systems almost always contain values that violate the documented schema. Then agree the target model and write an explicit field-level mapping, including the transformation rules and what happens to records that cannot be mapped. Build the extract, transform, and load as a repeatable, restartable job rather than a one-off script, so a failed run can be re-run without producing duplicates - that means idempotent loads keyed on a natural or surrogate identifier. Reconcile at every stage with control totals: source row count versus target row count, and a checksum on key numeric columns, plus a targeted sample compared field by field. Finally, plan the cutover itself - a full dry run against a copy of production, a defined freeze window, a rollback plan, and a period of parallel running where both systems are compared before the legacy system is retired. This open-ended question is testing structure, so state those phases explicitly before diving into detail.
Q: Write a SQL query to find duplicate rows in a table.
Group by the columns that are supposed to be unique together and keep only the groups with more than one row: SELECT email, COUNT(1) AS cnt FROM customers GROUP BY email HAVING COUNT(1) > 1; To see the full duplicate rows rather than just the offending values, join that result back to the table, or use a window function: SELECT * FROM (SELECT c.*, ROW_NUMBER() OVER (PARTITION BY email ORDER BY created_at) AS rn FROM customers c) t WHERE rn > 1; That second form is also the deletion pattern - keeping row number 1 as the survivor and deleting the rest - which is what a data-quality or migration cleanup task actually needs. Mention that NULLs do not group together as equal in some engines, which is a common source of missed duplicates.
Frequently asked questions about KPMG interviews
Section titled “Frequently asked questions about KPMG interviews”What is the KPMG interview process for freshers?
For KPMG Global Services (KGS, KPMG’s technology and delivery arm and the entry point for most CS/IT campus hires), the process typically runs 2-4 stages: 1. Online Aptitude Test (60 minutes) - five timed sections: Mathematics Ability, Logical Reasoning, Verbal Ability, Pseudocode, and Puzzles, with no navigating back once a section starts. 2. Group Discussion at some campuses (20-30 minutes) - current affairs or business topic. 3. Technical Interview (30-45 minutes) - projects, data structures, algorithmic problem-solving, and language-specific trivia. 4. HR Interview (20-30 minutes) - motivation, consulting fit, and a 2-year plan. Total duration: 2-3 weeks.
What questions are asked in KPMG interviews?
The OA’s Pseudocode section tests algorithmic reading rather than live coding, alongside math, logical reasoning, verbal, and puzzle sections. Technical interviews cover resume projects, data-structure fundamentals, language-specific trivia (memory management, paradigms), and how you break down an ambiguous problem. HR checks a clear 2-year plan, a genuine growth area, and comfort with client-facing work.
How many rounds are there in the KPMG interview?
KPMG typically runs an Online Aptitude Test, an optional Group Discussion, a Technical Interview, and an HR Interview - 2-4 stages depending on the drive. Some reports describe a tighter “2-round” campus process (test, then a combined technical+HR round); others run all four separately.
How should I prepare for KPMG interviews?
Three focused weeks across SQL basics, data interpretation, and pseudocode-style algorithmic reading is enough for most KGS drives. Revise data structures and language-specific trivia (memory management, paradigms), prepare one crisp project narrative, and practise structuring answers to ambiguous, open-ended problems - KPMG interviewers explicitly test that.
What is KGS, and is the interview different from Audit or Tax roles?
KGS (KPMG Global Services) is KPMG’s technology and delivery arm - it’s what this page’s process describes: an OA with a Pseudocode section, a data-structures-and-project technical interview, and HR. KPMG’s core Audit, Tax, and Assurance hiring runs a separate, largely non-coding process built around aptitude, group discussion, and HR, aimed mainly at commerce/accounting graduates.
What is the Pseudocode section in KPMG’s aptitude test?
It’s one of five timed sections in KPMG’s OA, alongside Mathematics Ability, Logical Reasoning, Verbal Ability, and Puzzles. It tests your ability to read and trace through algorithm-like pseudocode and predict output or identify errors - closer to reading comprehension for code than to writing a program from scratch.