Claude Code for Python Developers:
Hands-On Agentic Coding Course
Claude Code for Python Developers: Hands-On Agentic Coding Course
2 Days • Live Course
Stop copy-pasting from ChatGPT. Learn to build Python projects with an AI agent that lives inside your codebase.
Hosted By
Real Python core team member. Uses Claude Code daily to build tools, automate workflows, and ship Python projects.
Last Tuesday, a Python dev named Marco spent 45 minutes arguing with ChatGPT about a bug in his Flask app…
He pasted the traceback. It gave him a fix.
The fix broke something in a file it didn’t know about.
He pasted that error in—
It apologized, rewrote the function from scratch, and ignored the last three messages.
He tried explaining the project structure. It lost context halfway through. He started a new thread. Re-pasted his code. Got a different answer that contradicted the first one.
He closed the tab, opened his editor, and fixed it himself. It took twelve minutes.
Marco isn’t a bad developer. He ships code every week. He’s been writing Python for six years. He uses AI every day.
And every day, he has the same thought: there has to be a better way to do this.
He’s right. There is.
We recently surveyed 278 Python developers. Most of them sounded a lot like Marco:
“AI is good for solving a small task, but only this.”
“Going through a long session then hitting context limits and having to start a new thread.”
And this one hit a nerve:
“It’s so easy to lose track of an application’s architecture. There’s a very specific kind of frustration when something breaks, and you have no idea how to trace through the code.”
The pattern is always the same. Chat-based AI can’t see your project. It doesn’t know what files you have, can’t run your tests, and forgets everything when the context window fills up. So you end up doing the integration work yourself: copying code between windows, re-explaining your project, reviewing output you don’t fully trust.
65% of the developers in our survey are stuck at exactly this point. They use AI for code, it helps with small things, but it falls apart on anything real.
And underneath all of it, a question most developers won’t say out loud: if I can’t figure out how to work with these tools, what happens to my career?
There’s a different way to work.
What Changes With Agentic Coding
Instead of Copying Code From a Chat Window,
You Work With an Agent Inside Your Project
With Claude Code, the AI runs in your terminal. It reads your files, edits them directly, runs your tests, sees the errors, and fixes them. It manages git and works across your whole codebase, not one snippet at a time.
The first session of this course makes this concrete. You’ll start from an empty directory and watch the agent scaffold an entire Python project, then build features, run tests, and manage git, all from your terminal.
Over two days (4 hours each, via Zoom), you’ll build a complete Python CLI application from an empty directory using Claude Code. Not toy examples, but a real project with Click, Textual, uv, git, GitHub, tests, and multiple features. You’ll scaffold it, plan it, debug it, and ship it.
You’ll leave with the working project, a portable toolkit of reusable Claude Code skills, and a workflow you can apply to your own code the following Monday.
Course Curriculum
Both days are 4 hours of live, hands-on instruction via Zoom. You code along in real time. Q&A runs throughout.
Before the Course
Setup (We Help You Get Ready)
You’ll get step-by-step setup instructions several days before Day 1. When you sit down on Saturday, everything should just work:
- Claude Code installed and running
- Claude Pro (or Max) subscription active
- Python and uv ready
- GitHub CLI authenticated
If anything gives you trouble, we’ll help you sort it out before the course starts.
Saturday, March 21, 2026
Day 1 — From Zero to Working Project
You start with an empty directory. By the end of Day 1, you’ll have a working Python CLI app on GitHub, built entirely with Claude Code, plus a set of reusable skills you can take to any project.
Get Oriented: The Agent in Your Terminal
Before you write any code, you’ll get comfortable with Claude Code itself. You’ll learn how it sees your project, how to manage context and models, and why these matter more than you’d think. This isn’t a lecture, you’ll be running commands and seeing results in real time.
Outcome: You understand how the agent works and how to stay in control of context and costs.
Scaffold a Project From a Single Prompt
You’ll feed Claude Code a scaffolding prompt and watch a complete Python project appear: package structure, pyproject.toml, CLI entry point, logging, the works. Then you’ll run it. This is the moment chat-based AI can’t replicate: the agent creates files, wires them together, and gives you something that runs.
Outcome: A working Python project you can run with uv run second_brain.
Git & GitHub, Handled by the Agent
Instead of switching between your terminal and Claude Code, you’ll use them together. Initialize a repo, create commits, push to GitHub. You’ll see how permission modes let you decide how much autonomy to give the agent, and set up the feature workflow you’ll follow for the rest of the course.
Outcome: Your project is on GitHub with proper commit history and a repeatable workflow.
Set Project Conventions (and Make the Agent Follow Them)
You’ll curate a CLAUDE.md file that tells the agent how your project works: always write tests first, always add docstrings, always update docs before committing. From this point forward, the agent follows your rules on every feature, no separate TDD lecture needed.
Outcome: Project conventions that the agent enforces automatically.
Build Your First Feature End to End
This is where the workflow comes alive. You’ll plan a feature using plan mode, dictate the requirements (voice-to-text works great here), and watch the agent build it: code, tests, docs, PR. Then you’ll use the CLI you just built, and see Claude Code use it too.
Outcome: A working CLI with your first feature, built using the full plan, implement, verify, ship cycle.
Add More Features (or Take Them Home)
With the workflow established, you’ll run through it again with additional features. Depending on pace and group size, this is either hands-on practice in class or a set of well-defined GitHub issues you can work through on your own.
Outcome: More features shipped, or a clear backlog of issues ready for independent practice.
21
Day 1: Setup, Scaffold, Build & Ship
Sunday, March 22, 2026
Day 2 — Skills, Debug, Build & Ship
Day 2 shifts from building to maintaining and extending. You’ll create reusable skills, fix bugs, build a terminal UI, and leave with a toolkit that works on any project.
Create Reusable Skills (So You Never Repeat Yourself)
The feature workflow from Day 1 is powerful, but you don’t want to re-explain it every session. You’ll turn it into a portable Claude Code skill, type a slash command and the agent follows your workflow. You’ll also create skills for committing code and updating docs. These work on any project, not just this one.
Outcome: A Claude Code starter kit with skills you’ll keep using after the course.
Find and Fix What’s Broken
The project from Day 1 has bugs. (AI-generated code always does.) You’ll use Claude Code to systematically investigate the app, leverage your logs, reproduce issues, and fix them. This is the answer to the survey respondent who said “accepting code without a proper review causes anxiety.” You’ll learn a debugging workflow that actually works.
Outcome: A debugged, working app and a method for catching AI mistakes.
Build a Terminal Dashboard
The big Day 2 build: a Textual-based terminal UI for the app. First, the agent refactors your code, extracting business logic from CLI commands into a shared module. Then it builds a dashboard on top: notes list, search, preview. You’ll see a polished terminal app materialize, and learn how agentic coding handles refactoring and new interfaces.
Outcome: A terminal dashboard that runs alongside your CLI, built on cleanly separated business logic.
What’s Ahead: Hooks, MCP, Headless Mode
A quick-fire tour of what’s possible once you’ve got the core workflow down: automating pre/post actions with hooks, extending Claude Code with MCP servers, running it headless in scripts. Each topic gets a brief demo and pointers to go deeper. You’ll leave knowing not just what you learned, but what’s next.
Outcome: A roadmap for continued learning and a set of resources to explore.
22
Day 2: Skills, Debug, Build TUI & Wrap-Up
Course Dates: March 21–22, 2026
Enroll Today →Build Real Projects With an AI Agent
Full refund within 14 days if you’re not satisfied.
Real Python Satisfaction Guarantee
This course is backed by Real Python's guarantee. You can receive a full refund within 14 days after the course ends, provided you meet the completion criteria in our refund policy.What You’ll Be Able to Do
- Start any Python project with Claude Code and have it running in minutes
- Use a repeatable workflow to plan, build, debug, and ship features from GitHub issues
- Review AI-generated code with confidence instead of anxiety
What You’ll Receive
- 2 live half-day sessions (4 hours each) via Zoom
- Direct access to the instructor throughout both days
- Hands-on project building from scratch
- Lifetime access to course materials and recordings
- Portable Claude Code starter kit with custom skills and templates
- Printable cheat sheets and reference handouts
Who This Course Is For
#1
Chat Ceiling Hitters who use ChatGPT or Claude for code but are frustrated by copy-pasting, context loss, and re-explaining their project every session
#2
Copilot Users Ready for More who get value from autocomplete but want an AI that can read their codebase, run tests, and work across files
#3
Overwhelmed Tool Shoppers who feel lost in the flood of AI coding tools and want someone to walk them through one that actually works
Who Should Not Take This Course?
This course assumes Python proficiency. If you’re still learning Python basics, this will move too fast. Also, this is not a “learn Python” course. The focus is on the agentic coding workflow, not the Python language itself.
Meet Your Instructor
Core member of the Real Python team. Philipp uses Claude Code daily to build internal tools, automate content workflows, and ship Python projects faster.
Philipp isn’t teaching this course from slides he read last week. He uses Claude Code as part of his daily workflow at Real Python, where he builds tools, creates tutorials and video courses, and manages complex Python projects with agentic coding.
Before joining Real Python, Philipp studied type development at the Academy of Fine Arts Leipzig and spent over a decade working as a font engineer, writing Python programs that powered production pipelines for clients like Spotify, Slack, and The New Yorker. He knows what it’s like to adopt a new tool into an existing professional workflow, and he’ll share what actually works (and what doesn’t). Philipp is also a freelance software engineer and co-host of the Django meetups in Berlin.
Why Learn With Real Python?
Real Python isn’t just another course provider. Since 2012, we’ve grown from a Kickstarter project into one of the largest Python learning platforms in the world, reaching over 1 million developers, data scientists, and machine learning engineers every month.
Our content goes through what very few Python resources can match:
- Expert technical review for accuracy
- Teaching specialist evaluation for learning effectiveness
- Professional editing for clarity
Our live courses bring this same rigorous quality to an interactive, instructor-led format. You’re not just buying a course. You’re joining a decade-strong community that’s helped countless developers go from beginners to professionals.
What Learners Say
About Real Python Live Courses
"Ever since we've been taking the course, I've really changed a lot of the way that I'm programming. Honestly, just the first two weeks of this course was worth it for me."
— Matt Thacker, Sr. Solutions Engineer at Eptura
"Trying to self-learn is super hard. I saw the opportunity for this kind of structured class and by somebody in the community that I trust. And it's like, yes, I need the accountability. I need to have times to be able to present or show or meet."
— Chris Kessler, Software Engineer at Sony Pictures
"My code reading skills are greatly improved and I feel that I can understand not just the syntax and mechanics, but what the intent of the programmer might be and why they chose to write the way they did. This goes above and beyond so many other courses and tutorials that show you 'how it works' because this course shows you 'why it was built this way'."
— Jerry Wilson, Technical Lead at AIM EMS Software
The Alternative
You can keep figuring it out on your own. Read the docs, watch YouTube videos, try things that don’t quite work, start over. You already know what that looks like. You’ve been doing it.
Six months from now, you’ll likely be using AI the same way you are today. And the developers who learned a structured agentic workflow will be building at a different speed, on real projects, with code they actually understand and trust.
Two days, one tool, and a workflow you can use on Monday:
Course Dates: March 21–22, 2026
Enroll Today →Build Real Projects With an AI Agent
Full refund within 14 days if you’re not satisfied.
Frequently Asked Questions
You should be comfortable with Python fundamentals including:
- Writing and calling functions
- Using basic data structures (lists, dictionaries)
- Working with modules, packages, and virtual environments
- Basic command-line / terminal usage
- Basic git and GitHub familiarity
No prior experience with Claude Code or AI coding tools is required. That’s what the course teaches!
You’ll receive detailed setup instructions several days before Day 1. The essentials are:
- Claude Code installed and working
- An active Claude Pro subscription (Max recommended for intensive sessions)
- Python with uv installed
- GitHub CLI (
gh) installed and authenticated
We’ll provide step-by-step setup guides and support if you run into issues. The instructor will also pace exercises so everything runs smoothly.
Yes! You’ll retain permanent access to:
- The course repository (all code, skills, and resources)
- Session recordings
- Cheat sheets and handouts
- The portable Claude Code starter kit
This allows you to revisit the material and continue building on what you learned.
Both days run from 17:00 to 21:00 UTC. That’s 4 hours per session.
- Day 1 (Saturday, March 21): 17:00–21:00 UTC
- Day 2 (Sunday, March 22): 17:00–21:00 UTC
In your local time zone:
- Day 1:
Sat, Mar 21 @ 17:00 UTC toSat, Mar 21 @ 21:00 UTC - Day 2:
Sun, Mar 22 @ 17:00 UTC toSun, Mar 22 @ 21:00 UTC
Yes! Simply select the number of seats you need on the booking page. Each team member will receive their own access to the course materials and recordings.
This is a one-time payment that covers both days of the course plus lifetime access to all materials. There are no recurring charges or hidden fees.
This course is backed by Real Python’s satisfaction guarantee. You can receive a full refund within 14 days after the course ends, provided you meet the completion criteria in our refund policy.
Have another question? Email us at info@realpython.com
Course Dates: March 21–22, 2026
Enroll Today →Build Real Projects With an AI Agent
Full refund within 14 days if you’re not satisfied.
