LatentView Analytics Data Scientist Interview Questions
& Process

There are two stages in the interview process: first, you'll meet with a panel, and second, you'll meet with the manager or a client. They are primarily assessing your coding abilities, but they seem to care more about whether your solution matches their specific coding style or approach rather than your general problem-solving skills.

The interview process involved multiple stages. The first round was an aptitude test with 45 questions and 65 minutes to complete, after which 54 candidates advanced. The second round was a technical interview covering basic OOPS and DBMS concepts, interests, projects, some puzzles, and SQL queries, with 24 candidates moving forward. The third round was another technical interview focusing on areas of interest, statistics basics, puzzles, and academics, narrowing it down to 12 candidates for the final round. The fourth round was an HR interview. Although it was rumored to be non-elimination, it was indeed an elimination round. My interview was short, about 10-15 minutes, as I was the last person interviewed. I was asked about my family background, myself, and other general topics.

This was a pretty bad interview experience. I don't mind not getting shortlisted, but the interviewer seemed confused. For example, they expected me to discuss regularization when I was talking about model evaluation metrics, which doesn't make sense because regularization is for preventing overfitting, not evaluating. They also seemed to be looking for a very specific answer. In data science, solutions often depend on the situation, so there isn't always one right answer. I suggest having panelists who are experts in the roles they're interviewing for, as this reflects the company's quality. I'd also like for the panelist to be reviewed internally. I've never had to report an interview experience like this before, but I felt I had no other choice.