Generative AI and Large Language Models

Understand the structure, limitations, and applications of LLMs.

Pillar

Technology – Platforms, Tools, Infrastructure & Productivity

Overview

This course dives deep into the architecture and functionality of Large Language Models (LLMs), the foundational technology behind Generative AI. Participants will explore how LLMs are trained, their capabilities, inherent limitations, and practical applications across industries. The workshop also covers ethical considerations and strategies to optimize LLM use in business environments.

Learning Objectives

Participants will be able to:

• Explain the basic architecture and training processes of LLMs

• Identify common strengths and limitations of LLMs in various tasks

• Evaluate use cases where LLMs add value to business operations

• Understand challenges related to bias, hallucinations, and data privacy

• Apply best practices for deploying and interacting with LLM-powered solutions

Target Audience

• AI developers and data scientists

• Product managers and business strategists

• Innovation teams and technical leaders

• Compliance and risk management professionals

Duration

20 hours over 4 days (5 hours per day)

Delivery Format

• Instructor-led lectures with technical and business perspectives

• Case studies showcasing LLM applications

• Hands-on labs exploring LLM query design and evaluation

• Group discussions on ethical and operational challenges

Materials Provided

• Detailed course slides and reference guides

• Sample datasets and prompt templates for experimentation

• Checklist for ethical and responsible LLM use

• Access to LLM sandbox environments for practical exercises

Outcomes

• Solid understanding of how LLMs function and their potential impact

• Ability to critically assess when and how to deploy LLM-based tools

• Awareness of risks and mitigation strategies in LLM applications

• Practical skills to craft effective queries and interpret LLM outputs

Outline / Content

Day 1: Foundations of Large Language Models

• Overview of AI and Generative AI technologies

• Architecture of LLMs: transformers, tokens, embeddings

• Training data, pretraining, and fine-tuning processes

Day 2: Capabilities and Limitations of LLMs

• Natural language understanding and generation abilities

• Common limitations: hallucinations, biases, context length

• Real-world examples and failure cases

Day 3: Practical Applications of LLMs

• Use cases in customer service, content creation, data analysis

• Customization and integration options

• Hands-on practice with querying and output evaluation

Day 4: Ethics, Governance, and Future Trends

• Responsible AI practices specific to LLMs

• Privacy, fairness, and security considerations

• Emerging research and next-generation LLM models