After a recruiter call, I had a technical screen with SQL and basic product analytics questions.
The onsite was four interviews: presentation on a past analytics project, SQL, data analysis question, and analytics study. In the data analysis question, I was given a dataset and asked to find insights while working with the interviewer. Overall, it was not technically too difficult and the process was very well organized. Everyone was professional and nice. I thought I did well, but I didn't get an offer and I'm not sure what I did wrong.
Questions d'entretien [1]
Question 1
-- Revenue is down 10%. What do you do to investigate?
J'ai postulé via la recommandation d'un employé. J'ai passé un entretien chez Airbnb (San Francisco, CA) en avr. 2018
Entretien
The interview was for the inference data science group. I'm a candidate with ~15 years of experience in statistics and data science, with a quantitative Ph.D. I was asked to complete a data challenge, which is not unusual among DS jobs, but was not appropriate for my level of experience and demonstrated ability as a coder and researcher. With a family and many personal responsibilities, doing free work in the form of a data challenge -- rather than showcasing a real-world project I've done in the past -- was (in my opinion) not a good use of my time, and not a good way to screen for this position. I did the data challenge, got asked to proceed to the next step, and then declined. I decided it was not a place I wanted to work if this was the way senior and experienced candidates were being screened (as though I was a new undergrad with an unproven track record of past code and work).
Questions d'entretien [1]
Question 1
Data challenge was a project to build a statistical model to identify drivers of bookings on Airbnb. Data file was messy, and very realistic. Problems were very unstructured, and I used a lot of my own theory and style to craft a solution and make business recommendations based on my work. Also asked about designing an A/B test to get more clarity on the issues if more than observational data were made available. Wrote up a ~8 page memo with my results. Project took me about 2 days on a weekend.