| Model | Purpose | |-------|---------| | taxonomy-in-v2 | Predict 3‑level category tree for Indian e‑com | | hsn-predict-v1 | HSN code (4/6/8 digit) from description | | attr-extract-indic | Extract brand, flavour, variant, size, pack type | | regional-norm | Normalize regional spelling variations (e.g. dahi / curd , aloo / potato ) | | packshot-ocr-indic | OCR + text‑region detection for Indian scripts |
The year was 2081, and the dusty plains of Uttar Pradesh shimmered under a brutal sun. In a crammed server farm on the outskirts of Lucknow, a low hum escalated into a sharp whir. A single rack of quantum-hybrid processors blinked to life, displaying a single line of text on an obsidian screen: intelli catalogue ml version 80 india
—the "ML" (Machine Learning) designation reflects the platform's ongoing integration of AI to enhance part identification and search efficiency. mahindra-ecat.com Core Concept: The Shift to AI-Driven Cataloging | Model | Purpose | |-------|---------| | taxonomy-in-v2
: OEMs can go live with the new system in as little as 7 days , thanks to streamlined integration tools that connect directly to existing ERP and DMS infrastructure. A single rack of quantum-hybrid processors blinked to
Its architect was a 58-year-old, chain-smoking mathematician named Dev Rathore, who had designed its core logic in a windowless IIT Delhi lab. When the government finally greenlit deployment, Dev stood alone in the Lucknow facility, watching data cascade down the screen.
