中文

CodeSky · Markham · AI project-based learning

AI Project Course: from using AI to building with it

A student building an app on a laptop in a classroom, with a teacher nearby and screens showing a design, code, and a phone app.

Your child uses Anthropic's Claude Code to build a real app or website of their own. They go from understanding the request and writing the code, to putting it online, and then showing it to family and classmates.

We teach children to read code first, then to direct AI. When the AI gets something wrong, they can see where the problem is and fix it themselves — they don't stop at copy and paste.

A real teacher leads a small group and watches every step. Book a free trial and see what your child can make in the first class.

Book a FREE trial Project-based learning at CodeSky

Class hours: Monday to Friday 5:00–8:00pm, Saturday and Sunday 1:00–8:00pm.

9833 Markham Rd Unit 2, Markham, ON L6E 0E5

The whole project

One project, from idea to launch

Your child works through all five stages of a complete project. A real teacher leads a small group and watches every step.

  1. 1

    Understand the request

    Work out what to build and who it is for: classmates, family, or players.

  2. 2

    Design

    Sketch the idea and break it into small features.

  3. 3

    Code with AI

    Your child explains clearly what they want, writes the code together with Claude Code, and reads what each part does.

  4. 4

    Test and fix bugs

    Try it, have classmates try it; when something breaks, find the cause and fix it.

  5. 5

    Launch and show

    Put it online, show it to family and classmates, then improve it from their feedback.

Same AI, different results

Coders vs non-coders using AI: what's the difference?

With the same AI, a child who can code doesn't just get an answer — they know how to get the project done.

Reads & fixes

AI + a child who can code

Reads the error → fixes it → launches

Only retries

AI + a child who can't code

Can't tell where the problem is, so keeps asking AI to retry

Illustrative scenarios, not a judgment of any child’s ability or a promise of learning outcomes.

AI + a child who can code
AI + a child who can't code
Breaking down the request
Splits it into clear, small features
Hard to say exactly what they want
Reading the code AI wrote
Understands what each part does
Can't read it, has to accept it as is
Judging the approach
Can tell if it makes sense, and changes it
No way to judge
When it errors or a feature is wrong
Finds the cause, tests, debugs, fixes
Can only ask AI to try again
The end result
Truly online, and keeps improving
Stuck at a first result

Without a coding foundation, a child can still have AI generate code and produce a first result. But when something errors, a feature is wrong, or they need to change it, they often cannot tell where the problem is, and can only ask the AI to try again.

A child with a coding foundation can break a request into clear parts, read the code the AI wrote, judge whether the approach makes sense, then test and debug to get the project truly online and keep improving it.

So learning to code in the AI era is not about typing faster than AI. It's about giving children the ability to direct AI, check the results, and take responsibility for their own work. That is the core of how we use Claude Code to teach complete project development.

Who is Anthropic? What is Claude Code? Why Claude?

Anthropic is a U.S. AI research company that puts AI safety first. The AI it builds is called Claude.

Claude Code is Anthropic's AI coding assistant. It can plan first, ask questions when something is unclear, then write code and run tests — like an engineer partner working with you on a project.

Why Claude? First, its way of working is a real engineer's process: think it through, then build, then verify. Second, what it can and cannot do is set by the person using it, which is a good fit for learning safely with a teacher. Third, the child has to read, check, and decide at every step, which matches our approach: learn to write code first, then use AI well.

Where classes happen

9833 Markham Rd Unit 2, Markham, ON L6E 0E5

Class hours: Monday to Friday 5:00–8:00pm, Saturday and Sunday 1:00–8:00pm.

Phone: (647) 505-1788 · Email: info@codesky.ca

Book a FREE trial

FAQ

Does my child need coding experience first?
We teach children to read code first, then to direct AI. Without a coding foundation, a child can still have AI generate code and produce a first result. But when something errors, a feature is wrong, or they need to change it, they often cannot tell where the problem is, and can only ask the AI to try again. A child who can read code can break a request into clear parts, check the AI's work, test and debug, and get a project truly online.

What will my child make in the AI Project Course?
Your child uses Anthropic's Claude Code to build a real app or website of their own. They go from understanding the request and writing the code, to putting it online, and then showing it to family and classmates.

Who is Anthropic, and what is Claude Code?
Anthropic is a U.S. AI research company that puts AI safety first. The AI it builds is called Claude. Claude Code is Anthropic's AI coding assistant. It can plan first, ask questions when something is unclear, then write code and run tests, like an engineer partner working with you on a project.

How can my child try a class?
Book a free trial at our Markham studio, 9833 Markham Rd Unit 2, Markham, ON L6E 0E5, and see what your child can make in the first class. Email info@codesky.ca.

Book a FREE trial