Python Developer Assessment Template

Which Python feature produces lazy, iterable output that doesn't materialize the whole sequence in memory?

• A.List comprehension
• B.Generator expression
• C.Dict comprehension
• D.Tuple unpacking

You have CPU-bound work. Which is the best choice to parallelize it across CPU cores?

• A.asyncio.gather
• B.threading.Thread
• C.concurrent.futures.ProcessPoolExecutor
• D.multiprocessing.Queue

Hint: GIL matters here.

Implement `batched(iterable, size)` that yields lists of length `size` from any iterable, with the last batch potentially shorter. Must be lazy — do NOT materialize the whole input.

Given a list of dicts `[{'user_id': ..., 'amount': ...}]`, write a function that returns total amount per user. Use idiomatic Python (no loops that could be replaced with stdlib).

Write an async function `fetch_urls(urls: list[str]) -> list[str]` that fetches all URLs concurrently with a max concurrency of 10. Use aiohttp. Return the text bodies in original URL order. Handle per-request timeout.

Design a Pydantic v2 schema for a Task that has: - id: UUID - title: str (1-100 chars) - tags: list[str] (max 10, each lowercase-only, non-empty) - due_at: datetime (must be future-dated on creation) - estimated_hours: float (> 0 and <= 200) Write one test case that demonstrates validation failure for each rule.

When should you reach for `dataclasses`, `TypedDict`, and `Pydantic BaseModel`? Give one concrete use case for each and explain the tradeoff.