Async and concurrency

Top-three Python backend interview topic. Read in order — each file assumes the previous ones.

Foundations

#	File	Covers
01	01_gil.md	what the GIL is, why it exists, what it costs
02	02_concurrency_vs_parallelism.md	the distinction, and which Python tool gives you which
03	03_threads_vs_processes.md	OS-level: address space, scheduling, cost
04	04_context_switch.md	what a switch actually costs, OS vs coroutine

Asyncio

#	File	Covers
05	05_async_python_coroutines.md	`async def`, awaitables, how a coroutine runs
06	06_event_loop_internals.md	selectors, yield points, uvloop, Python vs JS loops
07	07_asyncio_vs_threads.md	the comparison, with hybrid patterns
12	12_taskgroup_structured_concurrency.md	TaskGroup vs gather, and picking the right primitive
13	13_cancellation_timeouts_errors.md	`CancelledError`, `asyncio.timeout`, shield, `ExceptionGroup`
17	17_anyio_trio_structured.md	anyio/Trio, and why libraries target them

Choosing a model, and offloading

#	File	Covers
08	08_cpu_io_tasks.md	classifying the workload — the decision that drives everything
09	09_when_to_use_processes.md	when processes are the answer
14	14_run_in_executor_to_thread.md	`asyncio.to_thread`, executors, sync drivers in async code
15	15_concurrent_futures.md	the unified pool API
16	16_process_pool_executor.md	process pools, pickling, IPC costs

Coordination

#	File	Covers
10	10_synchronization_primitives.md	locks, semaphores, events, conditions — across all three models
11	11_asyncio_queue_vs_threading_queue.md	queues, backpressure, and not mixing the two

Runnable examples: _examples/.

The 2026 headline

Free-threaded CPython is officially supported as of Python 3.14 (PEP 779) - no longer experimental, though still an opt-in build. Python 3.14 also added stdlib subinterpreters (PEP 734), so “threads, async, or processes?” is now a four-way question. Covered in 01_gil.md.