Most people ignore the importance of how they communicate during the interview, thinking that they are evaluated on their ability to solve and code a problem. This is a crucial mistake, and might very well cost you the position.

You need to communicate well with your interviewer, and make them part of the interview process. In the end, if you give the impression that you’d be weird to work with, the interviewer will have an implicit bias when evaluating your performance.

Practical tips:

• Ask leading questions. One of the best things you can do during the interview. It helps the interviewer to follow you and it allows you to verify you’re on the right track.
  • When coming up with examples:
    • Do you think I missed an interesting case?
    • Are the examples enough?
  • When proposing solutions:
    • Do you think this is an acceptable solution?
    • Do you think we can do better?
  • When coding:
    • Does this look good to you?
    • Does this make sense?
    • Right?
• Briefly comment on every choice you make and why did you make it (recursion/iteration, DFS/BFS etc).
• Keep the big picture in mind, ask if it’s okay to use abstraction functions (e.g. get_neighbors(node)) and code up helper functions later if you have time.

Keep in mind:

• Use type hints for Python. It’s part of the style guide for most companies, and it shows you care about good coding practices.
• Is the code testable and maintainable? Discuss maintainability/speed trade-off
• Prefer writing pure functions, avoid mutating function parameters
• Avoid relying on mutating global variables
• Do input validation is_valid(input)
• Use abstraction to your advantage! get_neighbors(node)
• State explicitly when cutting corners