Building a RAG AI Search Engine with LangChain

• Developed full-stack Shiny web application with Python utilizing retrieval-augmented generation (RAG) pipeline for retrieving and displaying accurate and contextually relevant results from a document database

• Created RAG pipeline with LangChain using Unstructured to parse and preprocess documents of various file types, OpenAI to generate embeddings, and Qdrant as the vector database

• Utilized ChatGPT LLM and prompt engineering to summarize relevant documents and answer the search queries

• Part of the Lonely Octopus Bootcamp, where we collaborated in a team and delivered a proof of concept for LixCap, an economic and transaction advisory firm

Example Work: https://rag-ai-search-engine.shinyapps.io/searchcap/