[realcertify]

The best interview that I ever experienced was when I applied for a Software Engineer position at VMware. The interviewers brought a laptop with an IDE and a project with a bunch of source code (based on one of their real in-house projects), and the task was to understand the exising code base and add a few new features to it. A real everyday task, no crazy algorithms or other menthal gymnastics.

[conor-]

Assuming an interview is a typical 30-60 min time slot is this actually a realistic ask? It seems pretty crazy to 1) get up to speed on a codebase and context in such a short amount of time to 2) then be able to actually develop against the codebase in a meaningful way to add more than one feature to it.

Can you share any additional details about what the process actually looked like or what type of task you were actually performing in the interview? Size of the project (~LoC), tech stack, complexity of the feature, etc.

[realcertify]

2 hours slot was allocated for that task. There was an API that I had to use to perform some operations and that made the task easier, as there was no need to study the entire code. C was used as the programming language, with a few thousands lines of code total. I completed the task in 2 hours successfully. I liked the fact that this method required zero preparations from the candidates side, and tested the actual abilities to code, as opposed to solving abstract problems.

[macsparrow]

Can be a fraught approach; interviewees could (and do) assume their answers will be used for production.

[golergka]

Just pay him for the time then. Mane companies already do this for test assignments anyway.

[philstephenson]

Seems like a question you could simply ask the interviewers up front. Establishing expectations about the task is useful anyway.