Building and Evaluating Advanced RAG Applications

What You'll Learn

• Apply advanced retrieval methods like sentence-window and auto-merging to improve RAG pipeline performance.
• Streamline your RAG evaluation process with best practices for iterative improvement.
• Use the RAG triad metrics—Context Relevance, Groundedness, and Answer Relevance—to assess LLM responses.

About This Course

Retrieval Augmented Generation (RAG) has emerged as a powerful method to leverage large language models (LLMs) with proprietary data. In this course, you will explore ways to optimize your RAG pipeline, including advanced retrieval methods and metrics to evaluate relevance and truthfulness of responses.

• Understand and apply sentence-window retrieval and auto-merging retrieval for context coherence.
• Employ evaluation and experiment tracking to refine and improve RAG pipeline performance.
• Implement the RAG triad of metrics—Context Relevance, Groundedness, and Answer Relevance—to assess LLM outputs.

By the end of this course, you’ll have the skills to build and evaluate a robust RAG pipeline for your specific use case.

Course Outline

• Introduction
  Overview of RAG and its importance in integrating LLMs with proprietary data.
• Advanced RAG Pipeline
  Techniques to improve RAG pipeline performance with code examples and best practices.
• RAG Triad of Metrics
  Detailed exploration of Context Relevance, Groundedness, and Answer Relevance for RAG evaluation.
• Sentence-Window Retrieval
  Implementing sentence-window retrieval to enhance context retrieval in RAG.
• Auto-Merging Retrieval
  Applying auto-merging retrieval to create coherent and relevant contexts.
• Conclusion
  Recap of advanced RAG concepts and key takeaways for application in real-world settings.

Who Should Join?

This course is suitable for anyone with basic Python knowledge who wants to effectively employ the latest methods in Retrieval Augmented Generation (RAG).