J'ai postulé en ligne. Le processus a pris plus d'une semaine. J'ai passé un entretien chez Fractal (Bengaluru)
Entretien
2 ML questions in 90 minutes? Seriously? Along with 9 mcq’s and expectation is to get 95% accuracy in the unsupervised problem and 70% in the supervised problem. This is a joke. An ML problem takes hours together to be analysed and thought about and to be coded. Do the Data Scientists at Fractal at least know how the interviews are being conducted? Or can at-least half of them crack these 2 problems in 90 minutes? Or is it just the 3rd party(DoSelect) coding platform randomly does this without any thought process. Anyway this is definitely not how you do it. Either keep one question to be solved or increase the time frame of the exam. Not expected out of such a big organization.
Questions d'entretien [1]
Question 1
Supervised learning problem and k means clustering problem.
Good and Short interview process
3 round of interview
1 take home assignment
1 round technical
1 round HR
each interview last 30min
Interviewer were supportive and helping
Whole interview take 1 week
Mildly challenging and nice. I would probably reduce number of rounds, i had 4 of them. The first two were technical and third one was managerial. The last round was done for cultural fit.
J'ai postulé en ligne. J'ai passé un entretien chez Fractal (Gurgaon, Haryana) en sept. 2025
Entretien
2 technical rounds, in first round basic to medium questions were asked. There were some coding questions as well which you need to solve.
In second round, it was basically system design kind of round based on your projects
