Generative AI Engineering Training

Overview

Learn the basics of Generative Artificial Intelligence (AI), its applications, and the techniques used to develop and engineer these systems. This GenAI course teaches attendees how to build and evaluate Generative AI models for a variety of tasks such as text generation, image synthesis, and music composition.

This Generative AI course gives you a comprehensive overview of Generative AI Engineering, from fundamentals to application. You’ll learn how to integrate LLMs (Large Language Models) into your AI applications, as well as the necessary steps to ensure your applications are secure and private.

Objectives

Understand the basics of generative AI and its applications
Learn about different techniques and algorithms used in generative AI
Develop skills to design and implement generative AI models
Gain proficiency in evaluating and optimizing generative AI models
Apply generative AI models to real-world problems

Duration

5 days

Who Should Take This Course

Audience

Programmers, Software Engineers, Computer Scientists, Data Scientists, Data Engineers, and Data Analysts.

Prerequisites

Extensive prior Python development experience
Core Python Data Science skills, including the use of NumPy and Pandas
Inferential statistics

Schedule

Course Outline

Introduction to Generative AI
- EXERCISE 1: Tools and Accounts Setup (existing exercise)
- Generative AI’s Roots in Machine Learning
- Understanding Generative models
  - DEMO 1: Generative ML
  - DEMO 1 Run Through: Students step through demo and make changes
- Contrasting Generative and Discriminative Models
  - DEMO 2: LLM Queries
  - DEMO 2 Run Through: Students step through demo and make changes
- The original LLM models – from BERT to GPT
  - EXERCISE 2: Your Own LLM Queries
- Current Cloud- and Offline-Based LLM’s
  - EXERCISE 3: GenAI Applications (design lab)
Generative AI Architecture
- Variational Autoencoders (VAE)
  - DEMO 1: BERT Vectors
  - DEMO 1 Run Through: Students step through demo and make changes
  - EXERCISE 1: BERT Vectors for Sentiment Analysis
- Generative Adversarial Networks (GAN)
- Reinforcement Learning from Human Feedback (RLHF)
- Transformers
- Generative Pre-Trained Transformers (GPT)
  - DEMO 2: Calculating Embeddings (from GPT model)
  - DEMO 2 Run Through: Students step through demo and make changes
  - EXERCISE 2: Sentiment Analysis with GPT Vectors
Tuning Generative AI Models
- Building Generative AI Models
  - DEMO 1: Fine Tuning for Q&A
  - DEMO 1 Run Through: Students step through demo and make changes
- How Pre-Training Works
- Data Preparation and Preprocessing
- Fine Tuning Generative AI Models
  - EXERCISE 1: Fine Tuning Tools & Steps
  - DEMO 2: Fine Tuning for Classification
  - DEMO 2 Run Through: Students step through demo and make changes
- Formatting Data for LLM Fine Tuning
- Fine Tuning GPT
- Transfer learning Techniques
  - EXERCISE 2: Fine Tuning with Full Dataset
]Evaluation and Optimization of Generative AI Models
- Intro
  - EXERCISE 1: Creating a Scoring Framework
  - DEMO 1: Summarization with GPT LLM
  - DEMO 1 Run Through: Students step through demo and make changes
- Evaluating model performance
  - EXERCISE 2: Designing an Evaluation Framework (design lab)
- Common evaluation metrics for generative AI models
  - EXERCISE 3: Creating an Evaluation Framework
Building Generative AI Applications (part 1)
- Intro
  - DEMO 1: String/Jinja/Template style formatting
  - DEMO 1 Run Through: Students step through demo and make changes
- Application Design Building Blocks
- Use Cases of LLM Based Applications
- Prompt Engineering Basics
  - EXERCISE 1: Prompt Engineering Exercises
- Prompt Templates
  - EXERCISE 2: Prompt Templates
- RAG with Llama Index
  - DEMO 2: Loading Private Data with Llama Index
  - DEMO 2 Run Through: Students step through demo and make changes
  - EXERCISE 3: RAG with Llama Index
Case Studies and Real-World Applications
- Generative AI for Text
  - EXERCISE 1: RAG Performance with Different Indexes
  - DEMO/Group Exercise: Reverse Engineering ChatGPT
  - DEMO/Group Exercise: Reverse Engineering Bing Chat
- Generative AI for Media
  - DEMO/Group Exercise: Reverse Engineering DALLE 2
- Generative AI for Code
  - DEMO/Group Exercise: Reverse Engineering Jasper AI
  - EXERCISE 2: Designing a LLM GenAI Application (design lab)
Building Generative AI Applications (part 2)
- Customizing with Prompt Engineering
- Advanced Prompt Types
  - EXERCISE 1: Advanced Prompt Engineering
- Customizing with RAG
  - DEMO 1: Few Shot Learning with GPT LLM
  - DEMO 1 Run Through: Students step through demo and make changes
- Customizing with SYSTEM/CONTEXT Arguments and Prompt Templates
- Customizing with Fine Tuning
  - EXERCISE 2: Using a Fine Tuned Model
- Design Considerations and Tradeoffs for Customizing
- Tying It Together with LangChain
  - DEMO 2: Tying it Together with LangChain
  - DEMO 2 Run Through: Students step through demo and make changes
ChatBots
- Chat Bot Basics
  - DEMO 1: Building a ChatBot with GPT
  - DEMO 1 Run Through: Students step through demo and make changes
- Building LLM-Based Chat Bots
  - EXERCISE 1: Building a custom ChatBot by hand
  - DEMO 2: Using LangChain to build a GPT ChatBot
  - DEMO 2 Run Through: Students step through demo and make changes
  - EXERCISE 2: Building a custom ChatBot with LangChain
Security
- Security Risks with Generative AI
- Secure Software Development
  - DEMO 1: Using an Offline LLM (Free Dolly)
  - DEMO 1 Run Through: Students step through demo and make changes
- Connectivity
  - EXERCISE 1: Building a Summarization App with an Offline LLM’s
- Exploitation of AI Systems (Jailbreaks)
- Infrastructure Concerns
  - DEMO 2: Using an Offline LLM with Langchain (part 1)
  - DEMO 2 Run Through: Students step through demo and make changes
- System Vulnerabilities
  - DEMO 3: Using an Offline LLM with Langchain (part 2)
- DEMO 3 Run Through: Students step through demo and make changes
- Data Privacy and Leaks
- Malicious Use of AI
- Obscuring Data for Privacy and Security
  - DEMO 4: Building Chatbots with Offline LLMs
  - DEMO 4 Run Through: Students step through demo and make changes
  - Best Practices for Security with Generative AI in Enterprises
  - EXERCISE 2: Building Chatbots with Offline LLM’s
Future Directions in Generative AI Products and Model Development
- Best Practices, Limitations, other Considerations
  - DEMO 1: LLM Agents Demo
  - DEMO 1 Run Through: Students step through demo and make changes
- Future of Work
  - EXERCISE 1/FINAL PROJECT
- Future Evolution of Gen AI
  - DEMO 2: Nemo GuardRails Demo
  - DEMO 2 Run Through: Students step through demo and make changes
  - EXERCISE 2: Working with GuardRails

FAQs

Is there a discount available for current students?

UMBC students and alumni, as well as students who have previously taken a public training course with UMBC Training Centers are eligible for a 10% discount, capped at $250. Please provide a copy of your UMBC student ID or an unofficial transcript or the name of the UMBC Training Centers course you have completed. Asynchronous courses are excluded from this offer.

What is the cancellation and refund policy?

Student will receive a refund of paid registration fees only if UMBC Training Centers receives a notice of cancellation at least 10 business days prior to the class start date for classes or the exam date for exams.