Arguably one of the worst interview experiences.
The hiring manager interviews you first, and honestly this is the best way to assess proficiency for a DS role. Was provided a churn data set and then expected to work with the interviewer in developing a model in a notebook. The open-book pattern allowed the process to be a discussion rather than ask-and-answered type interview. I believed this would continue throughout the process, but it just got progressively worse.
Next was a grueling 5hr interview process in a single day with 1 Data Analyst, few MLE engineers and 2-4 Data Scientists.
MLE engineers asked a Python Class problem in a code testing website, which was ok, but unnecessary for a DS Role.
Data Analyst was a SQL question in a Google Shared Document. Didn't need to be a special session, finished the exercise in 15 minutes and awkwardly spoke for another 30 mins.
Then came the interview with ALL the Data Scientists. This is where the impression hit rock bottom. The first 15 minutes were spent on talking about their own educational background and graduate research from ivy league schools, and it had no bearing on the kind of work that the company does. It just came off as elitist.
One of the Data Scientists, was condescending and kept rolling their eyes at answers. They would not accept alternative explanations, kept interrupting and then proceeded to present the same solution.
I think they need to figure out how to streamline the interview process some more and not expect a senior level candidate to allot 5 hours at a stretch, especially when they are working full time.
Perguntas de entrevista [1]
Pergunta 1
Questions were the same as the other poster:
Bayesian Inference
Binomial Probability coupled with A/B testing
L1 vs L2
bias - variance tradeoff
Comparing some model performance curves etc.
Candidatei-me por meio de recrutador(a). Fui entrevistado pela Migo Money em nov. de 2020
Entrevista
The overall process was pretty negative in a rather comical way. It started with the company down leveling a role from Director to Manager w/o a significant change in the responsibilities of the role (both w/ 10 yoe). I wasn't very interested but I went through with the process. The initial call was a bit awkward but generally fine.
The tech interview was very different than I expected. Most of the questions that I was given were something that you might find a college-level statistics, probability course, or possibly a bootcamp. He wanted to focus on minutia, details, and textbook definitions. He was pretty excited to correct me when he felt that I was wrong but on two occasions it seemed like he corrected me and then gave the same answer.
I am not sure if Migo is trying to implement the Rooney Rule in Eng but the process felt a bit awkward, forced, and my tech interviewer was working some things out on his end.
Perguntas de entrevista [1]
Pergunta 1
In a 60 minute interview we covered:
* Bayesian probability problem
* Binomial probability problem
* Algorithmic details and definitions of lasso vs ridge regression
* L1 vs L2
* Algorithmic details for random forest
* as well as the implications of the design choices
* Product questions on how Migo can do X, Y, Z.