Generative AI and LLMOps: Building Blocks and Applications

Course details

As large language models are becoming increasingly prevalent in various industries, professionals need to understand how to build, fine-tune, and deploy these models effectively and responsibly. In this course, learn the foundations of building, fine-tuning, and deploying LLMs in real-world applications. Instructors Soham Chatterjee and Archana Vaidheeswaran start with an introduction to LLMs and their evolution in the AI landscape. They then dive into LLM architectures, show you how to fine tune strategies for custom tasks, explain why and how to compress LLMs, and, finally, cover important aspects of prompt engineering. Through the course, they offer a series of challenges and solutions so you can practice the lessons as you go.

Concepts

0. Introduction

01 - LLMs for industry
02 - Industry-specific LLMs

1. Introduction to LLMs and Their Applications

03 - Understanding LLMs and their evolution
04 - Real-world applications of LLMs

2. Diving Into LLM Architectures

05 - Overview of LLM architectures
06 - How LLMs process and generate text
07 - The building blocks of LLMs
08 - Using a simple LLM
09 - Challenge - LLM for sentiment analysis
10 - Solution - LLM for sentiment analysis

3. Fine-Tuning Strategies for Custom Tasks

11 - Introduction to fine-tuning for LLMs
12 - Step-by-step guide to fine-tuning LLMs
13 - Best practices for fine-tuning LLMs
14 - Challenge - Fine-tune a pre-trained LLM
15 - Solution - Fine-tune a pre-trained LLM

4. Compression Techniques for LLMs

16 - Why compress LLMs
17 - Introduction to quantization and pruning
18 - Hands-on - Implementing compression in LLMs
19 - Challenge - Quantize a LLM
20 - Solution - Quantize a LLM

5. Prompt Engineering for Effective LLM Communication

21 - What is prompt engineering
22 - Best practices for effective prompt engineering
23 - Types of prompt engineering
24 - Challenge - Prompting LLMs to generate text
25 - Solution - Prompting LLMs to generate text

Conclusion

26 - Next steps