What are the current techniques being employed to improve the performance of LLM-based systems? How is the industry shifting from post-training towards context engineering and multi-agent orchestration? This week on the show, Jodie Burchell, data scientist and Python Advocacy Team Lead at JetBrains, returns to discuss the current AI coding landscape.

In our last conversation, Jodie covered how LLMs were approaching the limits of scaling laws. This time, we recap last year’s big focus on reasoning models and a post-training method called “reinforcement learning from verifiable rewards” (RLVR). We also cover test-time compute, where models spend more time reasoning through steps and considering multiple approaches to solve a problem.

We touch on Agent Context Protocol (ACP), agent orchestration layers, and context engineering. We also share some concerns about the hype cycle, maintaining all that code being generated, and running local models.

Course Spotlight: Vector Databases and Embeddings With ChromaDB

Learn how to use ChromaDB, an open-source vector database, to store embeddings and give context to large language models in Python.

Topics:

Show Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

What goes into managing a major project? What techniques can you employ for a project that’s in crisis? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

We discuss an article by Ben Kuhn titled, “How I’ve Run Major Projects.” We dig into the skills required for project management, and provide advice for when projects fall into crisis. We cover how the field’s terminology has been updated. However, the time investment, sober communication, and planning still remain at the core of successful projects.

We also share other articles and projects from the Python community, including recent releases and announcements, exploring an IDE for data science development, using Python set comprehensions, making friendly classes, using atexit for cleanup, a high-performance caching library for Python written in Rust, and a curated list of awesome marimo things.

This episode is sponsored by PropelAuth.

Video Course Spotlight: Using Data Classes in Python

When using data classes, you don’t have to write boilerplate code to get proper initialization, representation, and comparisons for your objects.

Topics:

News:

Show Links:

Discussion:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

With the mountains of Python code that it’s possible to generate now, how’s your code review going? What are the limitations of human review, and where does machine review excel? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

We discuss a recent piece from Glyph titled, “What Is Code Review For?” We dig into the limitations of human review and where software tools like linters and formatters can help you. We cover the challenges developers and open-source maintainers face with the rise of LLM-generated code and pull requests.

We also share other articles and projects from the Python community, including a collection of recent releases and announcements, creating publication-ready tables from DataFrames, choosing the right Python task queue, mastering context managers, statically checking Python dicts for completeness, an open-source inventory management system, and an ORM-based backend for Django tasks.

This episode is sponsored by SerpApi.

Spotlight: Intermediate Python Deep Dive: Write Better Python and Build Better Systems

Master advanced patterns, OOP, and Pythonic design in eight weeks–with live expert guidance.

Topics:

News:

Show Links:

Discussion:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

How do you quickly get an understanding of what’s inside a new set of data? How can you share an exploratory data analysis with your team? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

We discuss a recent Real Python tutorial about YData Profiling. This library lets you quickly generate an exploratory data analysis (EDA) report with a few lines of code. The report provides column-level analysis, visualizations, and summary statistics that can be exported to HTML to share with others.

We cover an article by Trey Hunner about building your own Python comprehensions. Python includes list, dictionary, and set comprehensions. But what if you want to create ones for other collections, such as a tuple, frozenset, or a Counter?

We also share other articles and projects from the Python community, including a collection of recent releases and PEPs, using the Django ORM as a standalone module, a history of attempts to eliminate programmers, a discussion of bad managers, a modern Python project template, and a CLI to summarize code size by language.

This episode is sponsored by AgentField.

Course Spotlight: Understanding Python List Comprehensions

Python list comprehensions make it easy to create lists while performing sophisticated filtering, mapping, and conditional logic on their members. In this course, you’ll learn when to use list comprehensions in Python and how to create them effectively.

Topics:

News:

Show Links:

Discussion:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

What goes into creating the tutorials you read at Real Python? What are the steps in the editorial process, and who are the people behind the scenes? This week on the show, Real Python team members Martin Breuss, Brenda Weleschuk, and Philipp Acsany join us to discuss topic curation, review stages, and quality assurance.

We start by sharing the multiple roles our panel of guests perform across the editorial process. They describe the phases a tutorial passes through, including layers of reviews, from technical accuracy to educational effectiveness. We also discuss our editorial independence, external authors, and the continuous feedback loop with our readers.

This episode is sponsored by AgentField.

Spotlight: Claude Code Course: Stop Copy-Pasting From ChatGPT

Most Python developers use AI as fancy autocomplete. This 2-day course teaches you to build entire projects with an AI agent inside your codebase.

Topics:

Show Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

Do you have complex logic and unpredictable dependencies that make it hard to write reliable tests? How can you use Python’s mock object library to improve your tests? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

Christopher shares details about his recent Real Python video course, “Improving Your Tests With the Python Mock Object Library.” He describes how mocking in Python with unittest.mock allows you to simulate complex logic or unpredictable dependencies, such as responses from external services. He covers how the Mock class can imitate real objects, and the patch() function lets you temporarily substitute mocks for real objects in your tests.

We also share other articles and projects from the Python community, including a collection of recent releases, using open source AI at Wagtail, updates from the inaugural PyPI Support Specialist, a lightweight OS for microcontrollers in MicroPythonOS, why match-case is not necessarily switch-case for Python, thinking about time in programming, a TUI-based presentation tool for the terminal, and a tool to check Django projects for dead code.

This episode is sponsored by AgentField.

Course Spotlight: Improving Your Tests With the Python Mock Object Library

Master Python testing with unittest.mock. Create mock objects to tame complex logic and unpredictable dependencies.

Topics:

News:

Show Links:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

How can you move your MCP tools beyond plain text? How do you add interactive UI components directly inside chat conversations? This week on the show, Den Delimarsky from Anthropic joins us to discuss MCP Apps and interactive UIs in MCP.

Den is a member of the MCP Steering Committee and a Core Maintainer focusing on auth and security. He explains how MCP acts as a universal bridge, providing AI models with the real-time context they need. He shares insights on working with MCP Apps and moving beyond simple text to render web-based user interfaces directly in your chat window.

Den previously worked on GitHub Spec Kit, and we discuss taking advantage of spec-driven development and how to use “Skills” to remove the toil from your workflow. We also talk briefly about his podcast and his conversations about navigating careers in technology.

This episode is sponsored by AgentField.

Course Spotlight: Getting Started With Google Gemini CLI

Learn how to use Gemini CLI to bring Google’s AI-powered coding assistance into your terminal for faster code analysis, debugging, and fixes.

Topics:

Show Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

Would you like to learn how to work with LLMs locally on your own computer? How do you integrate your Python projects with a local model? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

We cover a recent Real Python step-by-step tutorial on installing local LLMs with Ollama and connecting them to Python. It begins by outlining the advantages this strategy offers, including reducing costs, improving privacy, and enabling offline-capable AI-powered apps. We talk through the steps of setting things up, generating text and code, and calling tools.

We also share other articles and projects from the Python community, including the 2026 Python Developers Survey, creating callable instances with Python’s .__call__(), creating maps and projections with GeoPandas, ending 15 years of subprocess polling, discussing backseat software, a retry library that classifies errors, and a peer-to-peer encrypted CLI chat project.

This episode is sponsored by Honeybadger.

Course Spotlight: Tips for Using the AI Coding Editor Cursor

Learn Cursor fast: Use AI-powered coding with agents, project-aware chat, and inline edits to supercharge your VS Code workflow.

Topics:

News:

Topics:

Discussion:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

What are ways to improve how you’re using GitHub? How can you collaborate more effectively and improve your technical writing? This week on the show, Adam Johnson is back to talk about his new book, “Boost Your GitHub DX: Tame the Octocat and Elevate Your Productivity”.

Adam has written a series of books about improving developer experience (DX). In this episode, we dig into his newest book, which focuses on GitHub and how to get the most out of its features—from settings and keyboard shortcuts to hidden tools, CLI commands, and the command palette.

Adam also shares insights on the best ways to communicate on the platform. We discuss the nuances of GitHub-Flavored Markdown (GFM), best practices for effective communication in open source, the importance of maintaining civility in issue reports, and why he included a glossary of acronyms to help developers decode common shorthand like LGTM and FTFY.

This episode is sponsored by Honeybadger.

Course Spotlight: Introduction to Git and GitHub for Python Developers

What is Git, what is GitHub, and what’s the difference? Learn the basics of Git and GitHub from the perspective of a Pythonista in this step-by-step video course.

Topics:

Show Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas

How do you create automated tests to check your code for degraded performance as data sizes increase? What are the new features in pandas 3.0? Christopher Trudeau is back on the show this week with another batch of PyCoder’s Weekly articles and projects.

Christopher digs into an article about building tests to make sure your software is fast, or at least doesn’t get slower as it scales. The piece focuses on testing Big-O scaling and its implications for algorithms.

We also discuss another article covering the top features in pandas 3.0, including the new dedicated string dtype, a cleaner way to perform column-based operations, and more predictable default copying behavior with Copy-on-Write.

We share several other articles and projects from the Python community, including a collection of recent releases and PEPs, a profiler for targeting individual functions, a quiz to test your Django knowledge, when to use each of the eight versions of UUID, the hard-to-swallow truths about being a software engineer, an offline reverse geocoding library, and a library for auto-generating CLIs from any Python object.

Our live Python cohorts start February 2, and we’re down to the last few seats. There are two tracks: Python for Beginners or Intermediate Deep Dive. Eight weeks of live instruction, small groups, and real accountability. Grab your seat at realpython.com/live.

This episode is sponsored by Honeybadger.

Course Spotlight: Intro to Object-Oriented Programming (OOP) in Python

Learn Python OOP fundamentals fast: master classes, objects, and constructors with hands-on lessons in this beginner-friendly video course.

Topics:

News:

Show Links:

Discussion:

Projects:

Additional Links:

Level up your Python skills with our expert-led courses:

Support the podcast & join our community of Pythonistas