Reported on-campus pattern - Analyst
Source: consolidated from public student write-ups of Capgemini’s 2025-batch on-campus and Superset drives. The single most repeated line across accounts: every stage eliminates, and clearing one section does not compensate for a weak one.
Round 1 - Online assessment battery
Reported as several separately-timed sections in one sitting:
Pseudocode / technical MCQs (~25 questions, ~25 minutes)
-
Around 15 pure pseudocode output-prediction questions plus ~10 general programming MCQs
-
Topics that recur: arrays, stacks and queues, bitwise and arithmetic operators, and pre- vs post-increment trip-ups
-
Rated medium to hard by candidates - harder than the aptitude sections, and the most common place people are cut
English proficiency (~30 questions, ~30 minutes)
-
Grammar, synonyms and antonyms, tense-correct fill-in-the-blanks
-
Rated easy by almost everyone
Game-based aptitude (~4 games, ~27 minutes)
-
Capgemini replaced the older logical-reasoning MCQ section with gamified tasks, and this is well established in recent cycles
-
Reported game types: deductive logical thinking, inductive logical thinking, grid challenge, motion challenge
-
Generic reasoning prep does not transfer directly here - do practice runs on the actual formats
Behavioural / psychometric (untimed, ~20-25 minutes)
-
Roughly 75 questions in some accounts. Answer consistently rather than strategically
Round 2 - Coding assessment
Reported as top-performer gated - not everyone who clears Round 1 gets here. One 2025 write-up is candid that the candidate did not qualify for this stage at all.
- Typically 2 problems; most mainstream languages supported
- Difficulty reported as easy to medium, focused on fundamentals rather than heavy DSA
- This round is what separates the higher-band offers from the base analyst track in several accounts
Round 3 - Communication assessment
A four-section spoken-English test. Reported tasks: read and repeat, listen and repeat, conversation comprehension, and a short speaking task on an assigned topic - typically 30 seconds to think and 45 seconds to speak, with no repeats.
Candidates warn that the audio plays once only, so attention matters as much as fluency. Treat this round as eliminative.
Round 4 - Technical interview (25-30 minutes)
Reported as resume-driven and fundamentals-heavy, not algorithmic. Actual reported questions cluster into:
-
Self-introduction and project walkthrough
-
Which languages you know, and language-specific follow-ups - C++ pointers, friend functions, writing a palindrome program in Python
-
SQL fundamentals: basic queries, primary vs foreign key,
UNIONvsUNION ALL -
OOP concepts explained in your own words
-
SDLC methodologies - waterfall vs agile
-
Anything listed on the resume, including tools and certifications - one candidate was questioned on Google Analytics purely because it appeared on their CV
The most repeated advice: know every line of your own resume. Candidates report this being the decisive factor far more often than DSA skill.
Key takeaways: Pseudocode is the hardest gate. Practise the game-based formats specifically. The interview rewards clear fundamentals and honest resume knowledge over algorithm depth.