Prompt Engineering

In this lab, you will learn what prompt engineering is, why it matters, and how to design effective prompts for language models. You’ll practice zero/one/few-shot prompting, chain-of-thought (for step-by-step tasks), and sampling controls to balance precision and creativity.

Learning objectives

Upon completion of this lab, you will be able to:

• Discuss what prompt engineering is and why it is important for working with language models.
• Identify and apply the key components of an effective prompt.
• Use chain-of-thought prompting to improve reasoning and step-by-step problem solving.
• Apply zero-shot, one-shot, and few-shot prompting to different tasks.
• Adjust temperature, top-p, and (where supported) top-k parameters to control creativity, randomness, and output diversity.
• Evaluate and refine prompts based on model responses to achieve desired outputs.

Intended audience

This course is designed for:

• Data Scientists
• Software Developers
• Machine Learning Engineers
• AI Engineers
• DevOps Engineers

Prerequisites

Completion of previous modules is highly recommended before attempting this lab.

Lab structure

Demo: Mastering Prompt Engineering with GPT-4o-mini
In this demo, you will explore how small wording changes can dramatically change LLM outputs. You will:
- Break down the components of a strong prompt (role, audience, task, constraints, format).
- Use zero-/one-/few-shot prompting to steer behaviour.
- Apply concise chain-of-thought prompting for step-by-step tasks.
- Tune sampling controls like temperature and top-p (and top-k where supported).

Intended learning outcomes:
- Craft clearer prompts using role, audience, task, constraints, and format.
- Use zero/one/few-shot examples to guide the model.
- Trigger brief step-wise reasoning when appropriate.
- Adjust temperature/top-p to make outputs more precise or more creative.

Activity: Designing the Perfect Virtual Travel Guide
In this activity, you will improve a virtual travel-guide chatbot by tightening prompt structure, steering style with examples, adding light reasoning where needed, and tuning sampling for creativity vs precision.

Intended learning outcomes:
- Identify weaknesses in vague prompts and improve them with structure.
- Apply zero-shot, one-shot, and few-shot examples to guide behaviour.
- Use brief chain-of-thought prompting for multi-step questions.
- Adjust creativity with temperature