Secrets of Prompt Engineering

How to Communicate with AI Models

AI Assistants & Tools4.5(60)3h 21m30 lectures5 sections

What you'll learn

  • Understand AI, its types and challenges
  • Get to know ChatGPT
  • Use Zero-Shot Prompting
  • Apply Few-Shot Prompting
  • Understand Chain-of-Thought and Self-Consistency
  • Write prompts that include role, context and task
  • Set the tone and format of the reply
  • Use summarizing, translation and classification prompts
  • Write generation and programming prompts

About this course

This course teaches prompt engineering, which means how to write the instructions and questions you give to AI models such as ChatGPT so you get the best result. It starts with an introduction to AI, its types and challenges, and getting to know ChatGPT, then moves on to prompt techniques such as Zero-Shot, Few-Shot, Chain-of-Thought and Self-Consistency. After that you learn the elements of a good prompt: the role, the context, the task, the tone of the reply and the format of the reply. Then you learn the types of prompts, including completion, summarizing, generation, dialogue, translation, classification, question answering and programming prompts, and you finish with hands-on parts up close. The course suits anyone who uses AI for study, work, marketing or writing. You do not need previous experience or any knowledge of programming, because the course starts from the basics. By the end you will be able to write clear, organized prompts, choose the right technique for each task, and get more accurate answers from AI models.

Expected outcomes

  • Create effective prompts to maximize the benefits of AI models.
  • Ability to use prompts to enhance performance in summarization, translation, and programming applications.
  • Develop skills to optimize prompts using advanced techniques.
  • Master the preparation of prompts suitable for various scenarios.

Learner reviews

These reviews are from learners who bought the content and completed at least 50% of it — there are no reviews from anyone who hasn't gone through the material.
4.5
★★★★★
60 reviews
5★
36
4★
22
3★
1
2★
0
1★
1
  • ★★★★★خالد ص.·منذ 1 شهرشاهد 53٪

    كورس جميل وشرح مبسط ينتاسب مع جميع الفئات

  • ★★★★★سارة أ.·منذ 2 شهرشاهد 100٪

    دورة مفيدة جدا اتمنى ان يتم تحديثها للتطورات الجديدة واضافة اوامر انشاء صور وفيديوهات

  • ★★★★★محمود ع.·منذ 3 شهرشاهد 50٪

    جيد جدا ولكن العيب فقط ان الفيديوهات اكتر من اللزوم وكان يقدر يقلل الفيديوهات عن كدا

  • ★★★★★نبيل ا.·منذ 4 شهرشاهد 100٪
  • ★★★★★جورج س.·منذ 4 شهرشاهد 100٪

    الدورد جيده و الشرح ممتاز و لكن بالتطبيق على البرامج بالتاريخ الحالى وجدت بعض المعلومات اختلفت جدا على مثال chat gpt الان 5.5 و بامكانه البحث على الانترنت و ليس داتا مخزنه فقط

5 / 9 تقييم

Course content

1

Prompt Engineering

6 lectures
  1. Course IntroductionThe instructor introduces himself and the goals of the prompt engineering course: communicating effectively with AI models through natural language and good prompt design, focused on ChatGPT, outlining its units and who it suits with no technical prerequisites.5:48
  2. Artificial Intelligence and Its CharacteristicsGives an introduction to artificial intelligence (AI) as imitation of human mental abilities such as learning, reasoning, and problem solving, explaining its reliance on natural language processing and generation, its applications, history, and key pioneers.6:04
  3. Types of Artificial IntelligenceExplains the classifications of AI: weak AI with limited responses and strong AI resembling the human mind, along with machine learning, natural language processing, image processing, and neural networks with their hidden layers.11:45
  4. Challenges Facing Artificial IntelligencePresents the challenges of AI, including data bias from training models on Western data, ethical issues and intellectual property rights, fear of job losses, and the need to take part in developing AI rather than only using it.11:47
  5. Introduction to ChatGPT Platform (Part One)Introduces ChatGPT as a chatbot built on natural language processing (NLP), explaining what GPT means, pre-training and fine-tuning, and how generative AI differs from a traditional chatbot with fixed responses.6:24
  6. Introduction to ChatGPT Platform (Part Two)Covers the evolution of ChatGPT and its add-ons such as plugins, the difference between GPT-4 and GPT-3.5, and its challenges: data bias, knowledge cutoff, hallucination, and ethical and legal considerations.9:21
2

General Introduction to Command Engineering

4 lectures
  1. Engineering Commands Up Close (Part One)Defines a prompt and prompt engineering as the field of wording instructions precisely to get good responses from generative AI models, explains next-word prediction, and lists well-known example models.15:23
  2. Engineering Commands Up Close (Part Two)Explains the concept of tokens and how the model automatically splits input text into language units, with a hands-on test of a Tokenizer tool to see how words, numbers, and emojis are split.6:39
  3. Engineering Commands Up Close (Part Three)Explains how to test and try prompts hands-on in OpenAI's Playground, how it differs from ChatGPT, how to pick the model, and how to adjust settings such as maximum tokens and randomness.11:06
  4. Engineering Commands Up Close (Part Four)Shows the importance of prompt engineering as a bridge between humans and AI models, best practices such as open questions and role-playing, the skills of a prompt engineer, and the field's challenges such as ambiguity and ethical and legal issues.10:55
3

Types of Commands

9 lectures
  1. General Introduction to Types of CommandsIntroduces prompt types and distinguishes direct prompts from example-based prompts, with a hands-on test using a capital-city question, then lists the main prompt types to be covered one by one.5:24
  2. Completion PromptsExplains Completion Prompts, which fill in a missing sentence or piece of information, tries them on examples such as the zakat threshold, and shows next-word probabilities and how completions vary between models.7:29
  3. Summarization PromptsExplains Summarization Prompts and where they are used, tests them on a long text using hashtag markers to separate the context, and shows how specifying the summary format, such as paragraph count, gives a more accurate result.4:06
  4. Generation PromptsExplains Generation Prompts, which make the model create new content such as creative text or a poem, tests them on different platforms, and shows that the maximum token limit can cut a reply off before it finishes.7:16
  5. Dialogue PromptsExplains Dialogue Prompts, where the model plays a role such as a customer service or HR employee to converse with the user, and tests them with examples like choosing a phone or planning a trip.5:02
  6. Translation PromptsExplains translation prompts for language models: direct translation between languages, and detecting the language of an unknown text before translating it, with hands-on tests and context-aware translation rather than word by word.3:20
  7. Classification PromptsExplains classification prompts that sort inputs into one or more categories, such as true or false, positive or negative sentiment, and article topics, stressing the need to give the model the text to be classified.4:31
  8. Answering Questions PromptsExplains question-answering prompts, from specific to open-ended questions, testing examples in the playground, and noting that earlier context can affect the answer while starting a new chat resets it.2:47
  9. Coding PromptsExplains coding prompts for language models: creating, editing, improving and reviewing code, testing a request for game code and then a code review where the model finds and fixes deliberately planted errors.4:06
4

Elements of Good Commands

6 lectures
  1. General Introduction to Good Command ElementsSummarizes best practices for writing prompts and introduces the five elements of a good prompt: role, context, task, tone of response, and response formatting, noting that the first three are essential.5:32
  2. RoleExplains the Role element of a good prompt, asking the model to play a persona such as a marketing expert or a poet, and tests how the style and focus of the reply change with the assigned role.5:37
  3. ContextExplains the Context element of a good prompt, meaning the information that defines what the request is about, and tests how the reply changes when the context changes, such as selling versus renting phones, or a poem praising versus satirizing Egypt.5:21
  4. TaskExplains the Task element of a good prompt, the core action to perform such as write, summarize, translate, or describe, and tests how the output changes when the task changes, such as a Facebook ad instead of a creative description.4:49
  5. Tone of ResponseExplains the optional Tone of Response element in prompt building, and how specifying a tone such as suspense, anger, sadness or love changes the model's answer to the same request, with practical examples.5:15
  6. Format of ResponseExplains the fifth and last element of a good prompt, the format of the response, and how to ask the model for output as a bullet list or a table, then recaps the five elements, required and optional.5:35
5

Command Techniques

5 lectures
  1. General Introduction to Command TechniquesIntroduces the four core prompting techniques: Zero Shot, Few Shot, Chain of Thought, and Self Consistency, explaining the idea behind each and an example of a logic problem that confused older models.4:05
  2. Zero-Shot Prompting TechniqueExplains Zero Shot Prompting, asking a direct question or instruction with no examples or extra information, and tries it on direct questions and on a prompt with a role, task, and context.5:35
  3. Few-Shot Prompting TechniqueExplains Few Shot Prompting, guiding the model with one example (One Shot) or several in the prompt so it follows the same pattern, tests it with capitals and continents, and shows the difference between placing them in System versus User.4:04
  4. Chain-of-Thought Prompting TechniqueExplains Chain-of-Thought Prompting, guiding the model with a sequence of steps or ideas so it gives a complete, detailed reply, and tests it by writing a story about a pet on more than one platform.5:19
  5. Self-Consistency Prompting TechniqueExplains Self-Consistency Prompting, based on self-consistency and logical reasoning, and tests it on a word problem about ages in Playground, ChatGPT, and Gemini, showing how results differ by model.10:58

Instructor

Eng. Moamen Mohamed

Eng. Moamen Mohamed

Telecommunications engineer with a Master's in Information Security and a certified trainer from Microsoft and EC-Council.
103.3K views18 courses

This course is part of the diploma

Related courses