Course outline

Four units, from foundations to projects

Foundations → Prompt → Vibe Coding → project pathways. Students understand, produce and take away.

The path is clear: foundations → Prompt → Vibe Coding → project pathways. Students first see how large language models “read” language, then learn checkable prompting, drive Python projects with natural language, and finish with demonstrable work plus ethics awareness.

The outline is written for Hong Kong schools around understand · produce · take away. Each unit below lists a lead, detail and points so principals, coordinators and teachers can see the classroom structure.

AI trends and technical foundations

Why everyone talks about LLMs—and how they actually “read” language.

Students meet chat interfaces daily but rarely separate generative AI from traditional programmes. This unit clarifies the trend, then returns to tokens, training data and why hallucinations appear.

The goal is not jargon drills. Students should say in their own words what an LLM is good at, where it fails, and how to check outputs.

How generative AI differs from traditional programmes: rules versus probabilistic generation
Tokens, training data, hallucinations: why AI can be wrong and answers must not be trusted blindly
From ChatGPT to core large-language-model ideas: shared principles behind the interface

Prompt Engineering in practice

Ask in a way that actually helps—verifiable and reusable.

Many students dump a vague idea in one go and get thin, hard-to-fix output. This unit structures prompts so students define role and task first, then add constraints and examples.

Reuse matters: the same method supports writing, problem-solving and coding help, with practice in follow-ups instead of giving up after one ask.

Role, task, constraints, examples: a four-part prompt method that can be checked
Break fuzzy ideas into executable steps, then request results stage by stage
Common failures: vague instructions, too many asks at once, no follow-up—and how to fix them

Vibe Coding and AI programming

Turn ideas into Python projects with natural language—not rote syntax.

Vibe Coding is not “let AI write and walk away”. Students state the goal clearly, let the model draft, then read, verify and revise. They remain responsible for the work.

Through Python mini-projects—chatbots, text classification or practical tools—they practise reading code and fixing bugs so natural language connects to runnable results.

Vibe Coding: state what to build, let AI draft, then you verify
Python mini-project directions: chatbots, text classification, practical tools
Read code and fix bugs—do not paste blindly; outputs need human checks

Projects, showcases and pathways

Take work away for competitions, school showcases, further study and ethics.

The first three units build foundations, prompting and coding habits. This unit turns them into a personal or group LLM application project students can demo—not a chat log.

AI ethics—bias, privacy, copyright and responsible use—frames showcase and competition directions with clear awareness.

Personal or group LLM application projects: ideas into demonstrable outcomes
Hong Kong Science and Technology Fair / school STEAM Day directions aimed at showcase work
AI ethics: bias, privacy, copyright, responsible use—think before you deploy

How the four units connect

Unit 1 builds shared language: how LLMs differ from traditional programmes and where hallucinations come from. Without that, later prompting and coding become blind trial-and-error.

Unit 2 turns asking into a reusable skill. The four-part method makes requests checkable—steps, constraints and examples—preparing students for Vibe Coding.

Unit 3 connects natural language to Python work: state the goal, let the model draft, then read, fix bugs and verify. “Produce” means runnable, demonstrable results.

Unit 4 packages the work into a project and adds ethics. Outcomes can map to school STEAM Day or related showcases; students leave with a project and responsible-use awareness—not chat logs.

Request a tailored quote

Tailor sessions, level and projects for your school

Get a quote WhatsApp 5744 5000 Phone 2612 5789

Phone 2612 5789　·　Fax 2612 5787
WhatsApp 5744 5000　·　Email info@vibecode.hk
Website vibecode.hk