I applied online. I interviewed at Capital One in Mar 2021
Interview
Phone screen, take home data science challenge (binary classification problem) and virtual onsite(technical, statistical role play, case study, hiring manager). The one with hiring manager includes behavioral questions plus talk about your experience. Statistical role play interview is one of the hardest rounds followed by case study and technical. The technical round may consist of questions around the project you did, tools and tech used and hypothetical DS questions
Interview questions [1]
Question 1
The usual flight delay problem for statistical role play, take home challenge was for credit card transactions. Interpreting output of different regression vs classification models for statistical role play can be tricky and challenging.
I applied online. I interviewed at Capital One (New York, NY) in Jan 2021
Interview
DS challenge
Onsite interview ( 4 interviews)
1. Business interview (Case Analysis)
2. Roleplay (Data science and statistical questions)
3. Technical (Data science and statistical questions)
4. Technical (Data science and statistical questions)
Interview questions [1]
Question 1
Case study: Burger company improving its revenues
Roleplay: A long question with multiple sections about a machine learning project and designing its multiple steps
Technical: Lots of data science and machine learning concepts
I applied through a recruiter. I interviewed at Capital One in Nov 2020
Interview
Very straight forward, recruiter was helpful. Included a phone screening, then pretty much a big onsite day virtually. Behavioral, Role-Playing, and Technical Section. The technical section was not too clear on what they'd ask, and it was a little confusing to understand they were asking me about the specific tools I had used by describing a situation to me of where I'd use them.
Interview questions [1]
Question 1
Tell me about a time you taught yourself a new concept
What was the hardest project you've done?
RP: walk through the analysis, given either a regression output or tree. Explain to non-technical business stakeholder who only has knopwledge of elementary statistics
Technical: How would you collaborate with junior data scientists if you were a manager? If you were leaving the company what kinds of things would you include in documentation, and how would you ensure the performance of your model as it is monitored?