How We Built an AI Material Harmonization Engine for India’s Public Enterprises | SIH 2026

In this video, we present our Smart India Hackathon project: Material AI by Team AKTARA for Problem Statement ID 26099 (Theme: Smart Automation). Across India's Central Public Sector Enterprises (CPSEs)—such as IOCL, ONGC, NTPC, SAIL, and Coal India—inconsistent material coding leads to massive duplicate procurement and millions lost in fragmented tenders. Material AI solves this by deploying a hybrid NLP and deterministic rules architecture: Frontend: React.js + Vite + Lucide React Backend: Java 21 + Spring Boot + Spring Data JPA AI Engine: Python 3.13 + FastAPI + Sentence-BERT (all-MiniLM-L6-v2) + TF-IDF + FAISS for sub-second vector matching Database & Cloud: MySQL 8.4 hosted on Aiven Cloud, containerized with Docker, deployed on Render Watch our end-to-end live demo showing how different descriptions for the same stainless steel bolt across CPSEs are semantically matched with 96.74% similarity and unified into a single National Golden Code with human-in-the-loop validation! 🏆 Smart India Hackathon 2026 | Team AKTARA