Some engineering-related interview questions to ask a data scientist candidate.
- How comfortable are you at a Linux or OS X command line?
- Can you navigate around?
- What are pipes?
- Do you use source control?
- How comfortable are you with git in particular?
- Are you familiar with pull requests and code review?
- Do you know to avoid common mistakes like checking large input or result data sets into the source code repo?
- How do you ensure that your results are repeatable? How do you make sure other people can run your code reliably and easily?
- If you had to make a separate HTTP request to some upstream API for every ID in a list of 100,000 IDs, how would you approach that?
- What if a request to a second upstream API was required? How does this change if the second request depends on the results of the first?
- How might you approach the problem of sharing a report you've made to people on a regular basis?