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 advanced course on Prompt Engineering and Memory Management offers you a deep dive into techniques that enhance the performance and interaction of Large Language Models (LLMs). Starting with the basics of prompt engineering, you will explore a variety of advanced strategies, from few-shot to zero-shot and chain-of-thought prompting. As you progress, you’ll dive into context and memory management, learning how LLMs retain and utilize memory for more sophisticated interactions. The course’s hands-on projects help you apply each technique, ensuring that you not only understand the theory but also gain practical experience with real-world scenarios. The course also covers retrieval-augmented generation (RAG), a cutting-edge method that integrates external data retrieval with generative AI to enhance model responses. Throughout the modules, you'll engage in building and optimizing complex workflows, from setting up memory management for chatbots to constructing a complete RAG pipeline. You'll explore its integration into user interfaces, making the final product both functional and user-friendly. This course is ideal for intermediate to advanced learners with a background in AI or programming. It focuses on individuals interested in refining their skills in AI model optimization, particularly in the areas of prompt design, memory management, and RAG application development. By the end of the course, you will be able to implement advanced prompting techniques, manage context and memory in LLMs, develop a functional RAG pipeline, and integrate these systems into interactive applications.

Syllabus

Prompt Engineering: From Basics to Advanced

In this module, we will guide you through prompt engineering, starting from the basics to more advanced techniques. You'll experiment with various prompting methods like few-shot, zero-shot, and chain-of-thought prompting, gaining practical insights on how slight variations can drastically influence the quality of AI outputs.

Context and Memory Management in LLMs

In this module, we will explore the critical role of context and memory in LLM performance. You will dive into key concepts like context windows, token limits, and memory management techniques, followed by practical applications where you’ll build a chatbot with memory retention for enhanced interactions.

Logging in LLM Applications

In this module, we will cover the essential aspects of logging within LLM applications, from understanding its importance to setting it up throughout the application lifecycle. You’ll gain practical experience by building a chatbot that incorporates effective logging practices to streamline debugging and system tracking.

Understanding Retrieval-Augmented Generation (RAG)

In this module, we will introduce you to Retrieval-Augmented Generation (RAG) and how it optimizes LLM performance. You’ll dive into RAG’s core components and gain insights into its practical applications, challenges, and how it differs from traditional generative models.

RAG PDF Workflow and UI Integration

In this module, we will focus on creating a RAG pipeline specifically for document-related tasks, guiding you through the setup of the embedding model and UI integration. You will create a seamless workflow that includes building, showcasing, and interacting with your RAG system through an intuitive interface.