Python developer screening questions for recruiters

Python covers very different jobs, so the first screen should confirm which kind of Python work the candidate has done (web back ends, data pipelines, automation or machine learning) and whether it matches the role. Fifteen to twenty minutes is enough to tell which one a candidate has really done.

What does a Python developer do?

A Python developer writes and maintains software in Python. Depending on the team, that means web services built with Django, Flask or FastAPI, scripts that automate business tasks, data pipelines, or machine learning code. The Bureau of Labor Statistics counts them among software developers, whose duties start with analyzing users' needs and then designing and developing software to meet them.

Check the version too. The Python Developer's Guide lists Python 2.7 as end of life since 1 January 2020 and Python 3.10 since 1 October 2026, so recent work should be on 3.11 or later.

Which skills should you check at each level?

What screening questions should you ask a Python developer?

These are examples we wrote for this page. Replace the framework names with the team's own.

What did you use Python for in your last role, and what did your code do day to day?

This sorts web, data, automation and machine learning experience in one answer.

Strong answer: one clear description, for example "I maintained a FastAPI service that receives order updates from partners and writes them to Postgres through SQLAlchemy."

Weak answer: "Scripting, data analysis, web development and ML," all at once, with no project.

How do you set up a Python project so a new teammate can run it on day one?

This shows whether they've worked on shared projects or only in their own notebooks.

Strong answer: a virtual environment (venv, Poetry, uv or conda), a pinned requirements or lock file, a set Python version, a README, sometimes a Docker setup.

Weak answer: "I just pip install what I need."

Your endpoint takes five seconds to respond. Where do you look first?

This shows whether they measure before they change things.

Strong answer: time each step first, then check the database queries (an ORM firing one query per row is common), slow calls to other services, and a profiler such as cProfile. They fix one thing and measure again.

Weak answer: "Rewrite it in a faster language" or "add more servers."

When did choosing a list instead of a dictionary or set cause you a problem?

This is a good junior question: it checks that they know the basic data structures from real use.

Strong answer: a story such as checking whether items were in a long list inside a loop, which was slow, and switching to a set or dictionary for fast lookups.

Weak answer: correct definitions with no example.

How do you test code that calls an external API or a database?

This shows whether they test code that depends on other systems, or only run it and hope.

Strong answer: pytest with fixtures, mocking the outside call (unittest.mock or a library like responses), a test database or container for integration tests, and a clear line between the two kinds.

Weak answer: "I test it manually with Postman."

When would you use threads, multiprocessing or asyncio?

Ask this of mid-level and senior candidates; it tests how they handle many tasks at once in Python.

Strong answer: asyncio or threads for work that waits on the network or disk, multiprocessing for heavy calculation, because of Python's global interpreter lock (the GIL). They give an example from their own code.

Weak answer: "Threads make Python faster," with no mention of waiting versus calculating.

Tell me about Python code you inherited that was hard to work with. What did you change?

This shows judgment and whether they improve code in small, safe steps.

Strong answer: added tests before changing anything, introduced type hints and a linter such as ruff, broke up long functions, released in small pieces.

Weak answer: "I rewrote the whole thing from scratch."

What red flags should you note?

• Django, Flask, FastAPI, pandas, TensorFlow and PyTorch all listed as expert skills, with no project described in detail.
• Only Python 2 experience for a role on a current codebase, or no idea which version they used.
• "Machine learning experience" that turns out to be one course project. Note it as a mismatch if the role is ML-heavy.
• Can't explain how they install and pin dependencies.
• Long pauses before each answer followed by a polished paragraph. Our guide to AI cheating in interviews covers what a recruiter can and can't see.

Ask about any of these in your own call. If identity is the concern, read how to spot a proxy interview.

How do you run this screen with Simplehire?

Add the skills from your hiring manager call to the role as required skills, for example Python, FastAPI, PostgreSQL, pytest and Docker. Name the framework as well as the language, because Python covers very different jobs. Then send the interview link. The candidate talks with our AI interviewer in the browser, and the screening report is yours to read before you submit anyone.

Basic ($2 per screen, about 15 minutes) covers role-based questions and suits first screens on junior or high-volume Python roles. For candidates you plan to submit, Adaptive ($3, about 20 minutes) adds follow-up questions that dig deeper into the candidate's answers, plus a per-answer AI assessment. Advanced ($5, about 30 minutes) goes deeper still for senior or client-critical roles and adds an optional coding round, scored separately from the conversation.

The report gives a verdict, Skills demonstrated X/Y and Areas to probe. It is a recommendation, and the hiring decision is yours. Proctoring is on for every screen, and you can add an ID check for remote roles. You get 25 Basic screens free every month without a credit card (pricing).

If the role is mostly pipelines, the data engineer questions fit better. For a Python role that also owns infrastructure, see the DevOps engineer questions.