Build with LLMs: Prompt Engineering & Real AI Projects

Overview

This course features Coursera Coach! A smarter way to learn with interactive, real-time conversations that help you test your knowledge, challenge assumptions, and deepen your understanding as you progress through the course. This intermediate-level course dives deep into the essential techniques for working with large language models (LLMs), providing you with the practical tools and knowledge needed to build powerful AI applications. You will learn core concepts such as tokenization, log probabilities, and context windows, and gain hands-on experience through a series of engaging labs and projects. Additionally, you'll explore foundational prompt engineering patterns to improve response accuracy, control output formats, and develop advanced techniques such as few-shot prompting and chain-of-thought prompting. The course is structured to guide you from understanding the theoretical underpinnings of LLMs to applying those concepts in real-world scenarios. As you progress, you will work on an exciting project, building an AI-powered Git commit message generator. Through this project, you'll gain valuable experience in designing prompt templates, creating core logic for AI commits, and enhancing functionality with user reviews and model selection. Each section includes practical labs and exercises that ensure you're not just learning the theory, but also building real skills. The course is perfect for intermediate learners looking to enhance their AI development capabilities, particularly those interested in applying prompt engineering to optimize model behavior. A solid understanding of programming and LLMs is beneficial, but anyone with a technical background will find this course accessible. By the end of the course, you will be able to design structured prompts, implement advanced prompting techniques, generate AI-driven Git commit messages, and effectively manage API usage and costs.

Syllabus

Core Concepts: The Inner Workings of LLMs

In this module, we will dive into the key concepts behind large language models (LLMs), such as tokenization and log probabilities, and their impact on text generation. You'll also gain insights into the factors that contribute to API usage costs and learn how to optimize them. Practical labs will provide hands-on experience with these concepts.

Foundational Prompt Engineering Patterns

In this module, we will explore essential prompt engineering patterns that enhance the effectiveness of prompts. You'll learn how to structure prompts using delimiters, set personas, and implement few-shot examples to improve model accuracy. Additionally, we'll cover techniques like chain-of-thought reasoning and organizing information with templates.

Project Module #2: AI-Powered Git Commit Messages

In this module, we will guide you through creating an AI-powered tool for generating Git commit messages. You'll learn how to retrieve Git diffs, design prompt templates, and integrate user review steps. Testing and logging features will also be implemented to ensure the tool works effectively in real-world scenarios.