NLP Support Ticket Auto-Routing
Production-ready NLP system that automatically routes customer support tickets with 67.2% accuracy and 1.28ms latency.
Project Overview
An NLP-based system that automatically categorizes and routes incoming customer support tickets to the appropriate department, reducing manual triage time and improving response efficiency.
Technical Approach
Built with PyTorch for deep learning models and Flask for the API server. The system processes incoming tickets, classifies them using trained NLP models, and routes them in under 2ms latency.
Key Results
67.2% classification accuracy, 1.28ms average inference latency, production-ready API endpoint, significant reduction in manual routing effort.
Technologies Used
Python, PyTorch, NLP, Flask, Pandas, NumPy