Building Systems with LLMs - Syllabus

• Introduction to Building Systems with LLMs
  • Overview of using large language models for system automation
  • Introduction to building workflows with chain calls to the ChatGPT API
  • Course objectives and key concepts
• Language Models, Chat Format, and Tokenization
  • Understanding the chat format and token structure in LLMs
  • Basics of tokenization and how it impacts prompt engineering
  • Code examples demonstrating token use and limitations
• Classification Techniques
  • Classifying user queries with LLMs
  • Applications in chat agents and customer service automation
  • Hands-on code examples for implementing classification
• Moderation and Safety Checks
  • Evaluating user queries for safety and moderation
  • Ensuring appropriate responses through LLM moderation
  • Examples of handling sensitive or restricted content
• Chain of Thought Reasoning
  • Implementing chain-of-thought prompts for complex reasoning tasks
  • Guiding LLMs through multi-step reasoning processes
  • Code examples demonstrating chain-of-thought methods
• Chaining Prompts for Multi-step Systems
  • Building multi-step prompt chains for layered workflows
  • Interacting with completions of previous prompts to continue tasks
  • Examples of chaining prompts effectively
• Output Validation and Safety Checks
  • Methods to check and validate LLM outputs for relevance and accuracy
  • Ensuring outputs align with desired objectives
  • Code examples to implement validation checks
• Evaluation of System Performance
  • Evaluating the effectiveness and accuracy of LLM-driven systems
  • Assessing prompt performance for iterative improvement
• Project: Building a Customer Service Chatbot
  • Developing a functional chatbot using prompt engineering skills
  • Integrating classification, moderation, and multi-step reasoning techniques
• Course Conclusion
  • Summary of best practices and key takeaways
  • Final tips for responsible and effective LLM use in system automation