Candidates applying for Client Facing Data Scientist roles take an average of 45 days to get hired, when considering 2 user submitted interviews for this role. To compare, the hiring process at DataRobot overall takes an average of 32 days.
Common stages of the interview process at DataRobot as a Client Facing Data Scientist according to 2 Glassdoor interviews include:
Phone interview: 50%
Presentation: 50%
Here are the most commonly searched roles for interview reports -
I applied through a recruiter. The process took 2 months. I interviewed at DataRobot
Interview
I was connected through a recruiter. I had a brief, friendly HR phone screen, followed by several (>4) Google hangout interviews, mostly with Data Scientists. Everyone asked me to describe a predictive modeling project I had worked on in the past. Every interview focused on the technical aspects of the projects, even though the role I was applying for was presumably less technical and more about client management, which confused me. Then the communication completely stopped, and when I reached out to the HR contact I received the confusing reply that they were still searching but not in my area anymore. It was frustrating to put in an above average amount of time into an interview process and then get left hanging.
Interview questions [1]
Question 1
Tell me about a project you worked on. Why did you choose that particular target variable? Why did you choose that particular evaluation metric?
I applied through a staffing agency. I interviewed at DataRobot
Interview
I was first interviewed by a senior CFDS. The overall experience was quite good. But the second interview was a disaster.
In the second interview, I was interviewed by a junior (yes junior) staff.
The interviewer focused on jargons but not the concepts. For example, he asked me how to solve a particular problem and I tell him the workflow. He seemed not so happy and tell me a special jargon, and asked me did I know that thing. I said no and he explained the term, which is exactly the same workflow as mine. And I asked him so that was the same as what I had said right? He said yes that was the term.
Also he questioned a lot on my ways of building models, like why I did not use some complicated models and I said it is because of the regulatory requirement of my country. (Well I was using RF and boosting, but not NN, FYI) And he insisted he way of building models. (At least theoretically I think he was correct, but not applicable when there is a regulatory requirement)
He was not in the same country as me, but that's quite a shame to expand the business to another country without knowing their local law.
Finally, the 'model answer' he gave contradicts the constraints he gave me.
It is very unacceptable
Interview questions [1]
Question 1
My practical experience
What do you do if you are given a data set with no. columns > no. rows
I applied online. The process took 4 weeks. I interviewed at DataRobot
Interview
I got contact from the recruiter after one week of sending my CV. Initially she was very kind and showed a lot of interest in my profile. She scheduled a hangouts interview with a client facing data scientist. After the interview I had no feedback even though I asked for it several times to the recruiter.
Interview questions [1]
Question 1
Basic questions such as how random forest works, describe a project you did in your previous company, what kind of databases would you use for unstructured/structured data and how would you treat time-series.