Interview prep advice is mostly written for the US big-tech loop: five rounds, heavy algorithms, a system-design gauntlet. Most European Python roles are not that. They are shorter, more practical, and more interested in whether you can do the actual job. Here is the shape they usually take.
The four rounds you actually meet
- A screening chat (30–45 min). Mostly about fit, motivation and a quick sanity check on your stack. Easy to pass, easy to fail by being vague.
- A technical conversation or take-home. Either a 60–90 minute take-home with a clear time budget, or a live pair-programming session on a realistic problem. Pure whiteboard algorithms are less common for senior roles.
- A system / design discussion. "How would you build X?" They want to hear how you reason about data, APIs, queues and failure — not a memorised diagram.
- A values / team round. How you work, how you handle disagreement, how you communicate async.
The Python questions that keep coming up
You do not need to memorise trivia, but interviewers do probe whether you understand the language you claim to know:
- Mutable default arguments and why
def f(x=[])bites people. - The difference between a list and a generator, and when you would care.
- What
async/awaitactually buys you, and when it buys you nothing. - How the GIL affects threading versus multiprocessing.
- Where the ORM is generating a slow query and how you would find it.
None of these are gotchas. They are the things you hit in real work, which is exactly why they come up.
What interviewers are really listening for
- Do you reason out loud? A slightly wrong answer with clear thinking beats a right answer you cannot explain.
- Do you ask about constraints before designing? Jumping straight to a solution is a yellow flag.
- Can you talk about trade-offs without treating your favourite tool as a religion.
How to prepare without burning out
- Re-read the Python data model docs and actually run the tricky bits.
- Build or refresh one project end to end so you have concrete stories.
- Practise explaining a past decision in two minutes: the problem, the options, what you chose and why.
- Skip the 300-problem algorithm grind unless you are specifically targeting the few firms that test it.
The best preparation is having done the work and being able to talk about it honestly. When you are ready to put that to use, the open roles are here.