Demystifying Generative AI: A Practical Look Under the Hood
Published by O'Reilly Media, Inc.
Hands-on exercises to build AI agents, fine-tune models, and foresee biases in AI systems
Course Outcomes
- Be able to explain how a large language model (and generative AI) works in plain English and its strengths and weaknesses
- Have knowledge to build an AI agent using prompt engineering, temperature settings, RAG, and API integrations with external software.
- Understand bias, hallucinations, and other challenges in generative AI and manage them for safety.
Course Description
Generative AI is making industries rethink how we query and generate information, as well as interact with computers in general. Machines can generate text, images, and even code with remarkable human-like quality. Using Large Language Models (LLMs) and convolutions, the speed and rhetoric surrounding AI has never been more challenging to keep up with. This course for technical professionals demystifies how generative AI works under the hood, equipping learners with practical skills not just using generative AI effectively, but actually understanding its systemic strengths and limitations. Through interactive presentations and exercises, you will learn how to build AI agents, fine-tune models, and foresee biases in AI systems. You will walk away with the insights beyond just using generative AI. You do not have to learn to code these models to understand them, but rather have a common-language way of describing what they do.
What you’ll learn and how you can apply it
- Understand the fundamentals of Generative AI, including how Large Language Models (LLMs) and AI agents work.
- Develop a practical understanding of what makes LLMs and generative AI “tick”, including word embeddings and transformer models
- Gain insights with machine learning concepts and tools, and how to safely integrate AI into applications, projects or workflows.
This live event is for you because...
- You’re a machine learning practitioner, data scientist, or data engineer seeking a practical understanding of generative AI and how it works.
- You are a software developer or work with software teams wanting to understand the strengths and limitations of generative AI use cases.
- You are a programmer wanting to understand how to build generative AI tools and agents.
Prerequisites
- Exposure to machine learning principles
- Fundamental Python proficiency is helpful (variables, functions, loops, API calls, libraries)
Course Preparation
- Have access to a multi-modal chat agent like ChatGPT, Gemini, or Claude.
Course Follow-Up
- Hands-On Large Language Models (O’Reilly) - Jay Alammar, Maarten Grootendorst
- AI and ML for Coders - Laurence Moroney
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Generative AI Overview (25 minutes)
- Discussion: What is your excitement/concern about AI?
- Presentation: What is AI?
- Presentation: What is Generative AI?
- Presentation: What is a Large Language Model?
- Presentation: What is an AI Agent?
- Q&A
Machine Learning Basics (35 minutes)
- Presentation: What is machine learning?
- Presentation: Why is ML relevant in generative AI?
- Presentation: How machines “learn”
- Demonstration: PyTorch Linear regression/logistic regression
- Presentation: Deep learning and LLMs
BREAK 10 minutes
Large Language Models (50 minutes)
- Presentation: What are tokens?
- Presentation: What are embeddings?
- Presentation: What are transformers?
- Presentation: Predictions and temperature
- Demonstration: Unhinging an agent with temperature
- Exercise: Creating a jeopardy contestant bot
BREAK 10 minutes
AI Agents (25 minutes)
- Presentation: How LLMs are turned into actions
- Presentation: Retrieval augmented generation (RAG)
- Demonstration: Connecting AI to software
- Presentation: Prompt engineering
- Exercise: Social Media Politeness Bot
Images, Audio, and Video (30 minutes)
- Presentation: How AI makes images, video, and audio
- Presentation: Convolutions
- Presentation: Multi-modal AI
- Exercise: Guitarists and clocks
BREAK 10 minutes
AI Production Concerns (40 minutes)
- Presentation: When to use AI
- Presentation: Bias and Hallucination
- Presentation: Model validation
- Presentation: AI and job security
- Presentation: Selecting models
- Exercise: Choosing Sensitivity or specificity
Closing 5 minutes
Your Instructor
Thomas Nield
Thomas Nield is the founder of Nield Consulting Group and an instructor at O’Reilly Media and the University of Southern California, teaching classes on data analysis, machine learning, mathematical optimization, AI system safety, and practical artificial intelligence. He’s authored multiple books including Getting Started with SQL and Essential Math for Data Science, both for O’Reilly. He’s also the founder and inventor of Yawman Flight, a company that develops universal handheld controls for flight simulation and unmanned aerial vehicles. Thomas enjoys making technical content relatable and relevant to those unfamiliar with or intimidated by it.