Introduction
A machine learning system to classify and predict customer complaints for financial products, enabling quick identification of problematic complaints and helping companies take proactive action.
Approach
Use existing financial product complaints to train a model that identifies potentially problematic new complaints, helping the company take action to resolve issues.
Pipeline
- ML-based complaint classification
- Automated pipeline with Apache Airflow
- Docker packaging and CircleCI integration
Results
The four-month project delivered automated complaint triage. The original portfolio did not include numerical evaluation results.
Conclusion
Automating complaint classification supports more consistent triage and helps teams prioritize customer issues.
← Back to all projects