What changed in Python’s type system through 2024–2026 — PEP 695 type-parameter syntax, PEP 705 ReadOnly, TypedDict, Protocols, NewType, Self, and the patterns I keep reaching for in production.
Pydantic v2 in production. Model design, validators, serializers, settings, the strict mode, performance characteristics, FastAPI / SQLAlchemy patterns, and the migrations from v1.
Why Polars is the right DataFrame library for production Python in 2026 — lazy execution, Arrow-backed columnar engine, query optimization, the API differences from Pandas, and a pragmatic migration path.
Reliable structured output from LLMs in 2026. Pydantic AI, Instructor, OpenAI’s structured outputs / Anthropic tool calling, retry-on-validation patterns, and the end of homemade JSON parsing.
Picking a Python background job system in 2026. arq for async-native + Redis, Dramatiq for simple + reliable, Taskiq for typed tasks across brokers, Celery for legacy and giant ecosystems. With code, patterns, and a clear decision rule.
Modern asyncio in 2026. Structured concurrency with TaskGroup, ExceptionGroup, timeout/cancellation done right, anyio for cross-runtime portability, and the patterns to use (and avoid).
Modern Python tooling worth using in 2026 — uv replaces pip/venv/poetry/pyenv, Ruff replaces flake8/black/isort, ty (Astral) is the fast type checker, plus the project layout and pre-commit setup that pulls it together.
How async Django actually works in 2026 — async views, the async ORM, sync_to_async/async_to_sync gotchas, ASGI deployment with Daphne/Uvicorn, and where Channels still earns its keep.
A complete, opinionated production layout for FastAPI in 2026. Pydantic v2, async SQLAlchemy 2.0, Alembic, dependency injection, structured errors, settings, logging, and the project skeleton that survives the first 50 endpoints.
How to actually use the Anthropic Claude API in production. Messages format, tool use, prompt caching for 90% cost cuts, structured outputs, streaming, and the gotchas worth knowing.