WEEK 4
RAG with Citations
Answers from your docs and shows the source.
Course Details
Students build complete Retrieval-Augmented Generation (RAG) applications that generate answers grounded in retrieved documents instead of relying solely on the language model. They learn to retrieve relevant context, generate trustworthy responses with inline citations, measure answer grounding, and implement fallback behaviour when no reliable source is available, creating production-ready AI assistants that minimise hallucinations.
What you will learn
The full RAG pipeline end to end
Grounding answers in retrieved sources
Generating trustworthy inline citations
Measuring whether an answer is grounded
Fallback behaviour when no source fits
Topics Covered
RAG pipeline
Inline citations
Vector DB
OpenAI / Anthropic
Project 4
CitationRAG · DocuRAG
Project Details
Project 1: CitationRAG
Project Description
Build an AI assistant that answers questions from a document collection while providing precise inline citations for every response. Students implement the complete RAG pipeline, retrieve relevant document chunks, generate grounded answers, and display the exact source supporting each claim.
Project 2: DocuRAG
Project Description
Build an intelligent document Q&A system that enables users to upload documents and ask natural language questions. The application retrieves the most relevant context using the vector database, generates evidence-based answers, and gracefully handles situations where sufficient supporting information cannot be found.
Project Results
Project Result 1
Developed a citation-aware RAG assistant that retrieves relevant documents, generates grounded responses, and provides trustworthy inline citations for every answer. Implemented end-to-end retrieval, context injection, citation mapping, and answer validation to reduce hallucinations.
Project Result
Created a document intelligence platform that combines semantic retrieval with grounded AI responses. Built an end-to-end RAG workflow featuring context retrieval, answer generation, source verification, and fallback handling for unsupported queries, delivering reliable and explainable AI answers.
