One-Time Class: Advanced Prompt Engineering Masterclass for Teens

Most teens who use AI chatbots type a quick question and accept whatever comes back. This advanced masterclass shows what changes when you treat a prompt like a piece of engineering. In one focused 90-minute session, learners who already use AI tools learn the techniques developers rely on to get accurate, consistent and useful results from large language models.

We start by looking at how language models actually generate text, and why that explains common failures such as confident wrong answers, drifting instructions and vague output. With that model in mind, learners practise structured prompts with clear roles, context, constraints and output formats.

Next we move to advanced patterns: few-shot examples that teach the model a format, asking for step-by-step reasoning on hard problems, breaking big tasks into prompt chains, and requesting structured output like tables or JSON. Each technique is demonstrated live, then learners solve a short challenge with it and compare results.

The final part focuses on judgement. Learners design simple tests to check whether a prompt really works, spot hallucinations, and discuss responsible use, including academic honesty, privacy and bias. They leave with a personal prompt toolkit document they can keep using for coding, research and creative projects.

How classes work

The class alternates short live demos with hands-on prompt challenges. Learners work in their own AI tool, share results in chat and compare approaches with the group, with the teacher giving feedback on each attempt.

What we'll cover

• 

  Part 1: How Language Models Respond – Tokens, context and why prompts fail.

• 

  Part 2: Structured Prompts – Roles, context, constraints and output formats.

• 

  Part 3: Advanced Patterns – Few-shot examples, step-by-step reasoning, prompt chains and JSON output.

• 

  Part 4: Testing and Responsible Use – Evaluate prompts, catch hallucinations and use AI honestly and safely.

Learners should already use an AI chatbot regularly and be comfortable typing and editing longer prompts. Basic coding knowledge helps for the JSON example but is not required.