The Real Python Podcast

Christopher Trudeau, a Python educator and developer, returns to explore what happens to memory inside Python, from allocation and reference counting to garbage collection. He compares flat and src project layouts, packaging, testing, and entry points. The conversation also ventures into Parquet’s data-engineering dominance, Django upgrade planning, runtime benchmarking, and Pendulum’s delightfully cursed datetime operator.

Seth Larson, the Python Software Foundation’s Security Developer in Residence and urllib3 maintainer, explores the changing world of Python security. They discuss LLM-driven vulnerability discovery, the evolution of Python’s Security Response Team, and the role of PEPs in coordinating defense. The conversation also covers trusted publishing, SBOMs for hidden dependencies, CI/CD protection, cross-ecosystem collaboration, and Rust in CPython.

Christopher Trudeau, a Python developer and educator, returns for a lively tour of beginner pitfalls, from import shadowing and swallowed exceptions to mutation and surprising zip behavior. He also explores better development environments, safe temporary files, function design, Django decoupling, browser-based game creation, Rust-powered logging, and the latest Python releases, PEPs, and community elections.

Geir Arne Hjelle, a Python developer and educator, returns to unpack CodeMoo, his teaching tool for revealing how coding agents operate. They explore agent harnesses, tool-calling loops, sandboxing, and the growing importance of context and architecture. Geir Arne also discusses Grill Me, model choices, custom linting with Garuff, and preserving programming skills in the AI era.

Christopher Trudeau, a Python educator and developer, joins Bartosz Zaczyński, a Python educator and technical writer, to explore Python 3.15’s upcoming features. They cover sentinel values, lazy imports, unpacking in comprehensions, frozendict, mathematical upgrades, and the Tachyon sampling profiler. They also discuss terminal improvements, free threading, the JIT, and when to upgrade.

Meagen Voss, Wagtail’s community manager, and Thibaud Colas, a core contributor shaping its technical direction, explore AI’s impact on open source. They discuss the surge of AI-assisted pull requests, responsible contribution policies, model transparency, environmental costs, and measuring real productivity. They also reveal Wagtail’s human-centered AI features, agent-ready 8.0 release, and renewed interest in the independent web.

Christopher Trudeau, a Python developer, educator, and technical course creator, joins the conversation on Django’s 2026 survey trends and how developers are using AI. They explore reproducible Python builds, PyPI security, CPython infrastructure, and the GIL debate. Also covered: Polars versus SQL, secure Django unsubscribe links, Pydantic physical units, and a tool that rewrites if chains as match statements.

Den Odell, a staff web engineer, author, and performance specialist with 25 years of experience, explores how to make software fast by default. They discuss performance budgets, profiling across the stack, library and architecture trade-offs, end-user pain points, CI safeguards, and whether AI-generated code performs well. They also look at Python, WebAssembly, and Den’s return to synthesizers.

Christopher Trudeau, software developer and curator of PyCoder's Weekly, returns to highlight Python news and projects. He compares complicated versus complex systems and practical patterns for tackling them. He explores repeatable data science pipelines, including DataFrame validation with pointblank and using configuration (Hydra) to avoid hardcoded paths. Other tools and community releases are also discussed.

Calvin Hendryx-Parker, AI practitioner and speaker on agentic AI orchestration, explains why large-context workflows lead to silent failures. He highlights context management, deterministic chunking and embeddings, audit trails and hooks. He also talks about stripping noisy file formats, spawning subagents with clean contexts, and using evals and CI to catch model drift.