Navigating the Documentation Maze: LLMs and RAG to the Rescue

Documentation is often not easily accessible for newcomers to a project, making it easy to search is a fundamental requirement to have a successful OS project. During this talk we'll discover how to build a RAG pipeline to use LLMs to solve this problem and ease discoverability in our documentation.

Transcript

1. Navigating the Documentation Maze: LLMs and RAG to the Rescue
6. DEMO TIME huggingface.co/spaces/deepset/search-all-the-docs
7. +
8. Indexing Document Store Converter Files Files Files Cleaner Splitter Writer
9. Querying Retriever Document Store Answer Builder Question Prompt Builder LLM Answer Sources
10. Indexing components
11. Indexing pipeline
12. Indexing run
13. Query components
14. Query pipeline
15. Query run
16. Q/A
17. @silvanocerza@mastodon.world silvanocerza/search-all-the-docs linkedin.com/in/silvanocerza discord.gg/9amgZMWJ haystack.deepset.ai