AI-powered assistant to help Prototyping Instructors (PIs, who staff the makerspaces) and users with quick access to highly relevant and useful information about the makerspace, tools, training requirements, policies, and protocols.
Research Focus: Smart Devices in Makerspaces
The AI PI project focuses on enhancing makerspace accessibility and operational efficiency through the deployment of a specialized, context-aware Retrieval-Augmented Generation (RAG) chatbot. The core research challenge is engineering a highly reliable, low-maintenance and low-cost system that can accurately and safely answer complex inquiries specific to the Georgia Tech Invention Studio’s equipment and procedures, thereby significantly augmenting the capacity of the Prototyping Instructors (PIs).
The primary objective is to develop a production-ready, secure, and easily maintainable RAG infrastructure. Key current projects center on developing a fully functional, secure Admin Dashboard which enables non-technical staff to easily manage the knowledge base. This includes integrating Azure Authentication for Georgia Tech accounts, implementing document upload/delete capabilities, and developing a robust architecture to ensure low-cost use. This infrastructure is estimated to save approximately 600 staff hours and over $10,200 annually in labor costs by offloading common repetitive user queries.
The system employs a multi-layered RAG architecture. User queries are initially routed through an OpenRouter LLM which classifies the intent into categories like Invention Studio, General Knowledge, or Invalid Use. Based on this result, the application takes the appropriate action: either querying the Pinecone Vector Database for relevant context, leveraging a base LLM for general information, or triggering a predefined safety warning. This hierarchical process is critical for ensuring both high accuracy and safety.
The long-term goal is to fully integrate the AI PI into the Invention Studio’s operational ecosystem as an indispensable resource. Future development includes implementing advanced features such as audio transcription for voice input, incorporating advanced analytics on user query patterns, and eventually scaling the solution across other campus makerspaces to maximize efficiency and further democratize access to essential making expertise across the university.
AI PI Core Chatbot Application
AI Admin Portal Log-In: Implements Georgia Tech Authentication with Microsoft Azure