Online tests begin
First test that I gave was of Zomato on 16 October.
Zomato — It consisted of 4 easy-hard questions to be solved in 1 hour, time constraints, and one digit DP question made this test to be on the difficult side.
Saavn — It consisted of 3 questions to be solved in 1 hour, one was easy, one was easy-medium, and 3rd was the lab assignment of 2 weeks given to be solved in 1-hour test :). No one was able to solve this question!!
Graviton
Graviton held 2 separate tests for quant and SWE profiles.
SWE test — It consisted of 2 coding questions and 4 systems questions on OS and COA to be done in 1.25 hours. I solved 2/4 system questions and 1.5/2 coding questions.
Quant test — It consisted of many mcq type quant questions and in the end 2 subjective questions, which were really good and interesting, I could solve only 1 subjective question during the test but was able to solve the 2nd one (one of the best questions on the linearity of expectations) later.
Infurnia — It was mainly an aptitude test and didn’t have any coding questions.
Plutus Research — It was a single test of both profiles. It had 4 coding questions of which I was able to solve 2.5. And 5 quant questions (which took a lot of my time) in 1.5 hours.
OnePlus — They had 2–3 coding questions. I was not able to solve one question fully, it was based on the exclusion — inclusion principle and involved a lot of maths.
MTX — It had around 70 MCQs based on systems, aptitude, and 2–3 coding questions of medium-hard level (one was based on segment tree).
JPMC — It had quant questions and 2 coding questions of medium level.
Google (31/10/2021)
It was the test for which I was waiting eagerly. The questions were really on the tough side. We were given 2 questions to be solved in 1 hour from the mixed bag of questions. I was able to solve 1.9/2 of them. I couldn’t find anyone who was able to completely solve 2 questions!! Google also gives a lot of weightage to kickstart ranks and I had several under 500 ranks which bore me good fruit.
Goldman Sachs — It had 2 easy-medium level questions, an aptitude section, system’s section, and essay section. They removed the hard-level problem’s section this year.
Millennium Management — It had 3 questions of medium — hard level and I almost took the complete time of 1.5 hours to solve these 3 questions. It was an interesting test
Quantbox — It had a common test for all the profiles and the test was really on the hard side. It took almost everything to perform well in this test.
Uber — It had 3 questions to be solved in 1.5 hours. One was easy, one medium and one hard. But I had already solved a similar question as the hard one before so I solved all 3 in one hour.
Quadeye and NK securities — both the companies’ coding questions were standard questions but the quant part had a lot of questions as compared to the time given.
APT Portfolio — It had only one question based on the system’s coding. We had to implement memory unit in 2 hours. I was busy the whole 2 hours but in the end, I was able to design a system by the use of classes. It was one of the most challenging tests.
Sprinklr — It was a tough one. There were 3 problems, one was easy-medium and among the 2 others each one would have at least 1800 rating on codeforces.
Apple — The test for Apple was on the easy side and almost everyone was able to solve all questions.
Alphagrep — It had questions on DP, trees and other complex topics, it was a perfect CP-based contest and was on a hard level. I was not able to perform well in this test. In one question I faced TLE because I was not passing the vector as a reference. So always remember passing STLs as reference.
Rubrik
One of the most difficult and interesting tests of all. It had 4 questions. One was based on lazy propagation in the segment tree, one was based on observation, one on DSU, and one on matching, of which I was able to solve half of the cases by applying a greedy approach. I was able to solve 3.5/4 questions, but as many were not able to solve the lazy propagation problem, people solving 2.5/4 were also shortlisted.
