Rag Demystified
summary
A side project becomes an AI-powered event recommendation engine written in Ruby and built as a RAG pipeline where retrieval is the hard part. The system uses an embedding model to convert text into vectors plus an LLM for generation, hits needle-in-haystack problems when relevant facts sit deep in documents, faces context-window versus latency tradeoffs if the whole internet is queried, and requires curated ranking and human review to keep answers accurate despite working end-to-end.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:14Self-directed conversation
- 4:39Career into AI engineering
- 9:36RAG pipelines with context windows
- 11:24Defining RAG
- 16:40Retrieval and ranking challenges
- 17:44Non-technical AI tool tips
- 20:16RAG code surprises
- 22:24AI integration future
- 28:27Embedding models
- 31:57Closing and Q&A
related talks
53mBuild RAG from Scratch
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
27mOptimizing RAG Systems in Practice: Challenges and Proven Solutions
42mTo RAG or not to RAG
56mRAG: from dumb implementation to serious results
27mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
35mThe Hidden Jewels to Solve Confabulation Challenges in Generative AI’s RAG Promise
44mBuild RAG from Scratch