Python — Advanced
Engineer serious Python systems: the data model, metaprogramming, advanced typing, concurrency and async at scale, measured performance, architecture, security, and a production-grade capstone.
63 lessons0 challenges~26.5 hoursBeginner — no experience needed
Prerequisite course:Python — Intermediate
Who this course is for
Learners who finished Python — Intermediate and want to build, optimize, secure, and operate production-grade Python systems — making their own architecture decisions along the way.
What you will be able to do
- Implement objects that fully obey Python protocols via the data model, descriptors, and MRO
- Use metaclasses, __init_subclass__, and decorators for registries and plugin frameworks — and when not to
- Design type-safe libraries with generics, variance, ParamSpec, overloads, and Protocols
- Choose and benchmark concurrency models: threads, processes, asyncio, and the post-GIL landscape
- Build structured async systems with TaskGroup, timeouts, cancellation propagation, and bounded queues
- Profile with cProfile and timeit and deliver verified, measured optimizations
- Explain bytecode, frames, refcounting/GC, and the import system — and what they mean in practice
- Architect with ports and adapters, dependency inversion, repositories, and dependency injection
- Engineer production APIs: validation, error contracts, rate limiting, idempotency, pooling
- Build resilient job systems: retries with backoff, idempotency keys, circuit breakers, health checks
- Tune databases: query plans, indexes, isolation levels, N+1 elimination, optimistic locking
- Threat-model and repair injection, unsafe deserialization, SSRF, path traversal, and secrets leakage
- Test like an engineer: property-based, contract, and deterministic concurrency tests
- Instrument services with structured logs, metrics, correlation IDs, and health/readiness
The Data Model & Advanced Object Mechanics
Advanced Python starts under the hood: how attribute lookup works, what descriptors do to classes, how MRO resolves multiple inheritance, and building objects that behave like the built-ins.
Metaprogramming & Code That Writes Code
Decorators, __init_subclass__, metaclasses, and registries — with a professional bias for the simplest tool that works.
Advanced Typing & Static Design
Types as engineering tools: generics and variance, overloads, Protocols, TypedDict, ParamSpec — design first, checker second.
Concurrency & Parallelism
Threads, the GIL, executors, and synchronization — choosing a concurrency model deliberately and proving it correct.
Structured Async Python
TaskGroup, timeouts, cancellation propagation, async context managers and iterators — concurrency that composes.
Performance Engineering
Measure → profile → understand → optimize → verify. Call counts, algorithmic complexity, memoization, and generator streaming — with evidence, not folklore.
CPython Internals for Practitioners
Bytecode, frames, refcounting and GC, the import system — the mechanics that explain real Python behavior. Not a contributor course; a practitioner's X-ray.
Architecture: Boundaries That Earn Their Keep
Ports & adapters, dependency inversion, repositories, DI containers — every pattern justified by a problem it solves, plus refactoring a tangle without changing behavior.
Production APIs
Design HTTP APIs as contracts — status codes, envelopes, pagination, idempotency keys — then harden with layered validation, problem-details errors, and the production concerns that keep serving.
Distributed Systems Fundamentals
Why distributed systems fail: delivery semantics, idempotency, retry with backoff and jitter, at-least-once queues, dead letters, and circuit breakers under an injected clock.
Databases & Data Access
Relational thinking, SQL and transaction isolation, indexes and the N+1 problem, optimistic concurrency, connection pools, and repositories that speak the domain's language.
Security Engineering
Threat modeling before defense: injection in all its costumes, output encoding, authorization done right, password policy, secrets discipline, and ranking vulnerabilities in an audit.
Advanced Testing
The test pyramid and its budget, fakes over mocks, deterministic doubles, property-based testing with shrinking, flaky-test triage, and an honest test runner.
Packaging, Systems & Observability
pyproject.toml, wheels and lockfiles, semver as code, plugin registries, subprocess without footguns, atomic writes, graceful shutdown, structured logs with redaction, and error budgets.
Capstone: TaskFlow Platform
Build a production-style job platform from a requirements brief: exactly-once effects, retry budgets, graceful shutdown, observability, authorization — no tutorial, only acceptance tests.