Team17-Engineering Copilot using LLMs augmented with Public and Proprietary Documentation

Indian Institute of Science Summer 2025 DA225o Submission15 Authors

Keywords: Chatbot, RAG, LangChain, LLM

TL;DR: A chat-bot application capable of answering engineering queries in the field of semiconductor devices, built using publicly available Large Language Models (LLMs), software based on Retrieval Augmented Generation (RAG) and Langchain framework.

Abstract: This paper presents a chat-bot application that is capable of answering engineering queries in the field of semiconductor devices. It is targeted as a replacement to traditional engineer-to-engineer interactions for addressing customer queries on specific Integrated Circuits (ICs) by referring to company proprietary documentation such as IC datasheets, application notes, and user manuals, as well as publicly available information. This application is built using publicly available Large Language Models (LLMs), software based on Retrieval Augmented Generation (RAG) and LangChain framework. Performance comparisons of multiple RAG approaches are published based on a test dataset of questions and answers.

Submission Number: 15