Introduction to Large Language Models (LLMs) and Prompt Engineering by Pearson

Course details

Do you know how to launch LLMs like GPT, Llama, Claude, T5, and BERT at scale? This course provides a step-by-step approach for building and deploying LLMs, and it includes real-world case studies that illustrate the concepts. It covers how to begin your LLM journey with prompt engineering by showing how to make optimal instruction placements across models. And it explains how to build a Retrieval-Augmented Generation (RAG) system.

Whether you are a machine learning engineer, LLM developer, data scientist, or engineer who is interested in using LLMs for projects, this course is designed to help you get the best outputs from your models. It is helpful if you already have some Python 3 proficiency and have some experience working in interactive Python environments including Notebooks (examples: Jupyter, Google Colab, Kaggle Kernels, etc.).

Learning objectives
Build large language models (LLMs) that use semantic search.
List basic principles of prompt engineering.
Build agents and a retrieval-augmented generation (RAG) bot with OpenAI and GPT-4.
Describe how AI agents are built and operated.

Concepts

Introduction

Intro to large language models (LLMs) and prompt engineering

Overview of Large Language Models

Topics
What are large language models
Popular modern LLMs
Applications of LLMs

Semantic Search with LLMs

Topics
Introduction to semantic search
Building a semantic search system
Optimizing semantic search with cross-encoders and fine-tuning

First Steps with Prompt Engineering

Topics
Introduction to prompt engineering
Working with prompts across models
Building a retrieval-augmented generation bot with ChatGPT and GPT-4

Retrieval Augmented Generation + AI Agents

Topics
Introduction to retrieval-augmented generation (RAG)
Building a RAG bot
Using open source models with RAG
Expanding into AI agents

Advanced Prompt Engineering

Topics
Input output validation
Batch prompting + prompt chaining
Chain-of-thought prompting
Preventing prompt injection attacks
Assessing an LLM's encoded knowledge level

Conclusion

Summary