Data Science Interview and the Unicorns
Can you relate?
You read the bucket list of skills that typical data science jobs require or strongly prefer. You get discouraged. You sit down and make a list of the things and stuff you know and what you can pick up quickly. You read the job description again. You somehow decide that a subset of the skills can be safely ignored. You compare your notes with the job description again. Then you decide to apply. Whether the application will be glanced by a human, or at the very least, parsed by an intelligent system is a different story.
You get a call from a random recruiter. S/he found you via LinkedIn or some other site. The recruiter tells you a leading and fast-growing client is looking for a data scientist and you might be a good fit for that. Then you are asked about your years of experience in Python, machine learning, SQL, etc. If the recruiter is able to check the minimum number of required boxes, then the next question is how much you are making now and what you expect for a compensation. Then comes the decision whether s/he is going to submit your resume for the job.
None of these tells me what the job is about. Too much or too little information.
In the job search and interview process, if I am lucky enough to speak to a human being, I like to ask – “What does data science mean to you?”. I try to be subtle about it and phrase my question as nicely as my mood will allow.
Trust me, the recruiters and tech interviewers get baffled by this question more than I get baffled and drained by questions like
“How much real-work experience do you have as supposed to academia?”
“Tell me about your client-facing experience.”
“How do you manage a team?”
“Have you used deep learning?”
Well, I think all these questions are fair but they do not directly apply to all candidates. They don’t make much sense without a context either.
The whole point is to get to know the candidates and let the candidates know you. It should be a conversation not a series of questions.
I like to explain my projects to show the company the depth of the project and my understanding of the subject matter and approach to the problems. I am happy to go down the memory lane and recount the perks of the projects – difficult time, conflicts with a co-worker, quick decision making, choice of algorithm, strength and weakness, success and failure.
I would like my interviewers to do the same – explain a project they did for the company. That way I can ask questions and show them how much I get it.
No, not all recruiters or hiring managers are the same. The ones who can start and steer a conversation are the unicorns in the data science interviewing process.