Pooja K

BOM (Bill of Materials) – Text Extraction System

• Text Extraction System: Built a scalable PDF document intelligence pipeline using Python, Azure Form Recognizer, and Pandas. Implemented polygon-based table detection and OCR post-processing to extract structured BOM data.
• Reconstructed BOM hierarchies, including parts, quantities, and assemblies. Generated structured outputs (Excel and JSON) for enterprise integration.
• Automated extraction of engineering parameters and change history (ECN).

Deep Learning-Based Product Brand Identification System

• Developed an AI-powered warranty automation system using computer vision and Python.
• Automated verification of product images, invoices, and ownership documents.
• Reduced manual processing effort by 30%, and improved claim processing speed.
• Enabled real-time tracking and improved customer satisfaction by up to 90%.

NBFC Policy Compliance Automation Platform

• Built an AI-driven compliance platform to compare RBI regulations with internal policies.
• Automated regulatory monitoring, rule matching, and change detection.
• Centralized compliance data for improved governance and audit readiness.
  Reduced manual compliance tracking and minimized audit risks.