kb-python

Choose FastAPI for API-first services, Django for full-stack admin-heavy apps, and Flask for lightweight scripts. Framework selection depends on whether your workload prioritizes API design, built-in admin features, or minimal footprint.

Use async execution for I/O-bound concurrency to handle multiple waiting requests efficiently, and sync execution for CPU-bound work. Mixing them unsafely causes performance issues, so isolate CPU-heavy tasks to blocking threads or background queues.

Define typed public APIs with strong type hints and use Pydantic v2 for request/response validation and serialization. This boundary validation ensures consistent data shapes entering and exiting your Python backend services.

Select in-process background tasks for lightweight async workloads and distributed job queues for heavy CPU-bound processing. Structuring background workers properly prevents unsafe async and sync mixing while maintaining consistent domain error shapes.

Python 3.11 requires strong type hints and Pydantic v2 boundary validation for backend API design. These conventions ensure typed public APIs, consistent error response shapes, and structured request/response serialization across endpoints.

Structure testing with appropriate tools and async test patterns aligned to your execution strategy. Unit tests validate Pydantic schemas and logic, while integration and E2E tests verify endpoint behavior without unsafe async and sync mixing.