CQ_DataCrux | Sumit Sen | Fractal Analytics

Sumit Sen
Placed at Fractal Analytics

Interviewed By: Ankit

Hello Sumit! I am Ankit Ranjan from communiqué. First of all, a big congratulations for making it into Fractal Analytics. Please tell us about the general interview process for the companies you interviewed? Please mention the number of rounds and the nature of the interview process.

Thank you. There are three rounds of the interviews before which, they conduct a test to shortlist around 40 students. This year, 42 students were shortlisted based on it. There are a total of 62 questions in this test (2 hours test), four of which are coding problems and the rest are puzzles, math, or English questions. After you’ve been shortlisted, you’ll go through three rounds of interviews. The first two rounds can be interchanged. One round is related to business acumen to understand how much business understanding you have and the second is a tech-related interview. To the best of my understanding, there are two different types of interview processes. One is if you are into tech, and the other is if you don’t have much experience in tech. If you are not a techie, they will ask you some guesstimates or puzzles and will not get too much into tech; however, if you have a tech on your resume, they will focus more on the tech part, as was the case for me. In the first round, I was questioned about one of my internships at the Chicago Mercantile Exchange by a data analyst lead for one of the fractal’s teams. He was eager to understand how I learnt the financial concepts I needed during the project and how the tech implementation was done. They took around 15–20 minutes to discuss the CME project, and after that, they asked questions about my basic understanding of python. Then, I had a second round, after a break of 5–10 minutes. The second round was more of a tech. In this round, they didn’t ask standard DSA Questions, instead, they focused on deeper questions on machine learning at least this was the case for me. As Fractal has a lot of roles, they ask a variety of questions. For me, they took examples of ensemble learning algorithms and asked me about their working. So I had to give them a brief explanation of the algorithm, and the Mathematics behind it. Post this they quizzed me on a different project from my CV, which was my university of Iowa internship and then they asked me about what kind of data structures I had worked on, how did I implement that algorithm and what were things I was thinking of while implementing them. They were more interested to test my understanding of the math involved rather than the implementation of the algorithm. So, this was my second round of the interview. And my entire third round was devoted to discussing my reason for pursuing analytics after earning an undergraduate degree in Geology, as well as what I had done to prepare myself for the analytics sector.

Since you mentioned your project at the University of Iowa, I just wanted to know how significant is it to have a project from a reputed foreign university?

No, I don’t think that it is very important. As I told that Fractal’s interview process is more focused on the resume, right, and it depends on the interviewer as well. So in the first round, my CME group internship was more interesting for them. In the second round, it was my university of Iowa Project. So, it’s not specifically about the research internship that I am talking about but it’s more about the work that the interviewer is doing and how it is related to the work that I have done. And of course, adding a research internship and research project to your CV shows that you are more deeply into working on research-based projects, so they are expecting more of this from you. And since my project also focused on my ML work and which is why they asked a lot of questions on it.

Though you have already mentioned the types of questions they ask, I want to ask in particular if there are puzzles or more HR questions in the different rounds of the interview?

As I mentioned, my first round contained business understanding questions. Since my intern at CME required a lot of business understanding, my interviewer didn’t go into the puzzle and guestimate side of the interview process but one can easily expect guesstimates in the Fractal interview.

How did you prepare for the interview? Any valuable advice that you would like to share?

As previously stated, it’s more of a CV based interview, so you don’t have to prepare any different sorts of skills. They don’t expect you to know SQL or DSA, if you have several machine learning internships as well as PORs (as in my case), then you are good to go with Fractal, you don’t have to prepare specific technical aspects. However, you should have a solid understanding of machine learning algorithms and be familiar with guesstimates and puzzles. So you should know some basic case prep and machine learning aspects (for tech) and you should do more puzzles, guesstimates and case prep (for non-tech).

Since you mentioned the importance of CV in Fractal, can you suggest what should be mentioned in the CV to make it stand apart from the general public?

I’m not familiar with Fractal’s real shortlisting procedure. So, I’m not sure how they chose you. Most likely, my test score was the deciding factor in my selection. And for the things to be mentioned in my CV, like for the tech role, I had a corporate internship, a research internship and I do have several hackathons mentioned and apart from these tech-mentions, I do have several good PORs including that of a Hall President, Society Head and Governor. So, I would suggest trying to mention some PORs in your CV which is significant on the campus and apart from this try to put more tech into it. If you don’t have any tech experience, you can still get into Fractal by having a POR-focused CV and performing well in guesstimates. However, they do examine technical skills in your CV. So, try to include more tech in your CV, along with a few PORs that you believe are important on campus.

Can you suggest some strategies or resources that can be helpful for preparation for the tech domain?

For the tech aspect, there are two different things that I know. One is on the DSA part — you should be doing basic DSA beforehand. This is a compulsion for every interview process, whether it is fractal or any other company. I would recommend that you take any DSA course. There are various resources for learning DSA, thus I won’t go into detail about them. So, one tip from my side is that you should be well-versed in DSA. There are numerous domains in machine learning, such as natural language processing (NLP), computer vision, etc. In machine learning, I had an internship in statistical learning. I didn’t go into Deep Learning or Natural Language Processing, but I did mention it in every interview. In a machine learning interview, they want to know how much you know about the algorithm and mathematics behind it. So, for that, the first and the basic course that needs to be done is the Andrew NG course. Apart from that, I would suggest stat quest lectures on youtube. Statquest lectures are very beneficial and will give a better understanding of machine learning and its practical application. That is what they expect from you. Just do two internships or projects on machine learning and learn the basic course on machine learning. Like I did Udemy A-Z, Andrew NG course and apart from that, Statquest lectures and several hands-on projects on Kaggle.

As a pre-final year student, what can students do to maximize their chances of successful placement, especially if they have not done any internship till now and not got PPO?

For those pre-final year students two-three things I would suggest. But these solutions will help you if you are sure of going into tech. Firstly, be very thorough with your DSA prep. So, start practising DSA, it’s an important step for any internship and placement. If you are done with DSA then I am pretty sure that you would be landing up in a good job and I believe that 3 to 4 months of DSA would help you in getting a good placement. But you have to be very dedicated to it. Apart from that since my interest was in machine learning among all domains in tech, what I did was just one or two internships in machine learning and had a basic understanding of what machine learning is. So this is how I prepared for my placements and I would also recommend that if you want to go into analytics, then these are the things that will help you in getting a good job.

One more thing students preparing for the analytics domain are often unsure how much they should emphasize on doing competitive programming. Any advice on that?

The point is for the test you need DSA. Four coding questions are being asked in Fractal for the shortlisting process. So DSA knowledge is a must to clear it. DSA is not just about competitive programming, but it is the problem-solving skill that you have. So if you don’t have anything in your CV, then the interviewer needs to know how good of a problem-solver you are, and this is how they prefer to check. From my perspective, you can’t be completely into machine learning without knowing DSA. I am not asking you to do competitive programming on code forces and all but if you have done basic 400–500 questions from InterviewBit or LeetCode, then it would suffice. So if you are looking for an analytics domain, you have to do coding to get placed at good companies. So for getting into good companies like Fractal, you need to have a good knowledge of DSA.

Do you have any other suggestions for students preparing for placements this year?

I think there may be a lot of students facing the same problems during placement that I faced. I couldn’t decide which profile I should choose. So, I prepared for product management, analytics and coding as well. I was into various profiles. Hence I would suggest that you can follow any profile that you want but to secure your job then you should know how to code. If you want to take risks, you can directly work on product management or whatever you want to do. But if you want to get into a tech role, then coding is a must. Once coding is done then 80% of your work is done for placement. So according to me, coding is very important for getting into a good company if you are into tech or unsure about the profile to pursue.

Thank you so much Sumit for your time. It’ll surely be very helpful for all of the KGP junta preparing for analytics profile. Thanks a lot for giving this interview to Communiqué IIT Kharagpur and we wish you all the very best for all your future endeavours.

Thank you!

Back to list