DEV Community

Cover image for ChatGPT Prompt Engineering for Developers
Stefan Alfbo
Stefan Alfbo

Posted on

ChatGPT Prompt Engineering for Developers

I have recently finished the ChatGPT Prompt Engineering for Developers a course made by DeepLearning.AI.

This is one of their short courses that they offer for free.

The course has nine lessons where each lesson has a video about 10 minutes and a supplementary Jupyter notebook (Python) to aid the learning of the current topic. It's a great starting course if you want to know more about this topic.

The lessons are:

1. Introduction

Introduces the course and is led by experts Isa Fulford and Andrew Ng. The course will explore instruction-tuned LLMs.

2. Guidelines

Goes through two key principles of prompting. Key principles are clear instructions and giving the model time to think. Each principle has some tactics that involve delimiters, structured outputs, checking conditions, and few-shot prompting.

There is also a section about hallucinations and how to reduce them.

3. Iterative

Walks us through the iterative prompt development process.

Iterative prompt development

4. Summarizing

This lesson demonstrates how to use ChatGPT for text summarization. It covers summarizing reviews, modifying prompts for specific purposes, and extracting information. Also, it shows a workflow to summarize multiple reviews efficiently for better insights.

5. Inferring

Shows how to inferring with ChatGPT, where the model analyzes text to extract information like sentiment, topics, and emotions from reviews as an example.

6. Transforming

Here we explores the capabilities transforming text. Demonstrations of translation between languages, conversion of tones and formats, proofreading for grammar and spelling errors, and generating compelling and formatted content using various prompts.

7. Expanding

This lesson is the contrary to the Summarizing lesson. Expanding involves generating longer text from shorter prompts. It showcases personalized email generation, leveraging sentiment analysis and customization.

It also introduces the parameter temperature of LLMs to control response variety.

8. Chatbot

Walks through how to build a custom chatbot. The "OrderBot" example demonstrates building a conversational pizza order assistant.

Shows that customization and persona-building are possible by adjusting system messages and how to work with context.

9. Conclusion

A summary of the complete course.

Top comments (0)