Meta's internal recruiter reached out over LinkedIn.

The first stage was a phone screen with the recruiter (the same one who reached out). Fairly standard questions: "Tell me about your experience," "Walk me through your CV," etc.

Second stage was a coding round. Six questions in total:

• Three SQL questions: Nothing too difficult – JOINs, aggregations, filters. No more than three JOINs were required in a query. There were four tables provided and a schema. The trick here is not to rush; think about which tables you need and which ones you don’t.
• Three Python questions of varying difficulties. You start with a simple one-liner solution, moving to a medium question that requires loops, lists, and dictionaries. The last question is a hard one.

Tell me about your experience. Walk me through your CV.

What was the data that you worked with, what did you do with it, and how was it used?

You have a number with multiple digits. You need to construct the largest possible number from these digits.

You have a dictionary where keys are strings and values are lists of strings. You need to find the most frequently occurring string across all lists/values. You can’t use pandas or anything similar; use pure loops, lists, and dictionaries.

You have a list of data classes. Each data class represents an object with various start and end times of a meeting, as well as the number of people in the meeting. There are overlapping meetings. Your task is to find the maximum number of people that were in a meeting at once.