A chronicle of the MatSelect project: Building an AI-powered Steel Advisor and Alloy Designer during our industrial training journey.
MatSelect addresses the fragmentation of industrial steel data. Starting as a simple data collection task, it evolved into a full-stack Progressive Web App featuring an AI Steel Advisor powered by Gemini.
The project was developed by a team from Guru Nanak Dev Engineering College (Srijan & Chandan) under the mentorship of Dr. Khushboo Rakha at IIT Ropar.
Fast Frontend
Responsive UI
Intelligent Backend
Auth
An online kickoff meeting regarding project scope. Due to end-semester exams, I couldn't join offline immediately. Our team (Myself, Chandan, Nandini) from GNDEC was assigned the task of collecting steel data.
Presented the first prototype with 50 steel grades. Feedback received: Improve UI and add interactive elements. This led to the creation of the Sandbox for Material Exploration (later Alloy Designer).
Refined the UI significantly. On July 10th, Chandan and I joined Sabudh for a 6-month Data Science training, which limited our time availability but expanded our skill set.
Met Dr. Khushboo Rakha physically at DREAM Lab. We presented the PPTs featuring the AI Steel Advisor Chatbot.
Met with AI Expert Dr. Sugandh Rakha. He suggested critical features for the bot and guided us on improving the UX to make the interface more user-friendly.
A hectic period learning Vite + React while implementing. On Sept 2nd, Chandan and I deployed the prototype with Firebase Authentication (Google Sign-In).
Meeting with Sugandh Sir regarding UI modularity. Due to the Data Science internship, coordination became difficult. We are currently focusing on Sabudh but remain active members.
Next Phase: Continuing development in January 2026.