During my 5 years in architecture school, I was always struck by how little technology had permeated our daily workflows. With a background in computer science and a deep interest in programming, I was always curious to explore ways to enhance architectural processes through technology. As a passion project and learning exercise, I set out to build something of my own, a Retrieval-Augmented Generation (RAG)-powered chatbot trained on the Indian National Building Code (NBC) to help architects, engineers, and builders quickly navigate the complex world of code compliance with instant, accurate, citation-based responses.

This assistant is built using:

1. LangChain for orchestrating the RAG pipeline

2. Microsoft Azure Document Intelligence for parsing structured text and tables from the NBC

3. Chroma DB for embedding storage and retrieval

4. Customized opensource Chainlit for creating a lightweight, chat-based user interface