Python Interview Questions: From Fundamentals to System Design

Language internals (favourite of senior interviewers)

1. Difference between list, tuple, set, and dict — with time complexities for common ops.
2. How does Python's GIL affect multithreaded CPU-bound code, and when does asyncio help?
3. Mutable default arguments — why is `def f(x=[])` a bug, and what's the fix?
4. `is` vs `==` — when do small integers and interned strings mislead you?
5. Decorators: write a decorator that caches results with a TTL.
6. Generators vs iterators vs coroutines — when do you reach for each?
7. Context managers — write one with __enter__/__exit__ and one with @contextmanager.

Data structures & problem-solving

1. Two-sum in one pass using a hash map.
2. Find the first non-repeating character in a stream.
3. LRU cache with O(1) get/put — collections.OrderedDict or dict + doubly linked list.
4. Merge k sorted lists with heapq.
5. Detect a cycle in a linked list — Floyd's tortoise & hare.
6. Word ladder / BFS on strings.

Pandas & numpy (for data roles)

1. Difference between .loc, .iloc, and .at — when is each the right tool?
2. Group-by with multiple aggregations and rename columns cleanly.
3. Vectorise a row-wise operation that's currently a for-loop.
4. Merge two dataframes on multiple keys with different join types.
5. Handle timezone-aware timestamps and convert between UTC and local.
6. Efficiently read a 10 GB CSV that doesn't fit in memory (chunksize, dtypes, pyarrow).

Testing, packaging, tooling

• Write a pytest fixture with scope='module' and explain when you'd use scope='function'.
• Mocking: patch a network call in a function you don't own.
• Explain the difference between requirements.txt, pyproject.toml, and lock files.
• How does virtualenv differ from Docker, and when do you need both?

Applied system design

1. Design a rate limiter in Python — token bucket vs sliding window.
2. Design a background job system using Redis + RQ / Celery — retries, dead-letter, backoff.
3. Design an ETL pipeline that pulls from an API and lands in a warehouse — idempotency, watermarks, backfill.
4. Design a URL shortener service and describe how you'd deploy it (FastAPI + Postgres + Redis).