IssueSense AI
Turning support tickets into structured decisions with a ten-category NLP classifier.
The problem
Manual support-ticket triage takes time and makes consistent routing difficult.
The approach
Built a multi-class text classification pipeline using TF-IDF features and Logistic Regression to categorize tickets across 10 categories.
Results
The project reports 94.38% accuracy and a 94.4% macro F1 score on a dataset of more than 10,000 support tickets. These are project evaluation results, not production performance guarantees.
Integration
Serialized the trained model with pickle and created a standalone prediction script for integration into a ticketing workflow.