Severity-based profanity and inappropriate-content detection for high-volume user content.
- Role
- ML Systems & MLOps Engineer
- Section
- Data Projects
- Built with
- Transformers / Airflow / NLP / MLOps
How it works
Built a Thai-English profanity and inappropriate-content detection system for high-volume user-generated content. Moderation is graded by severity: highly unsafe content is banned automatically and borderline cases go to admin review. Batch processing, monitoring and retraining workflows orchestrated by Airflow.
