AI Engineer (Makita) — Nov 2025 to Apr 2026
Built an AI fraud-detection pipeline using XGBoost combined with deep embeddings trained in PyTorch, with results evaluated on precision-at-recall. Implemented automated feature monitoring to reduce drift and improve model reliability for production deployment. Led governance activities for production ML features consumed by enterprise customers. • Engineered training/feature workflows for fraud models • Applied evaluation logic (precision-at-recall) to track model quality • Used monitoring and experimentation tools to support model lifecycle • Coordinated design reviews and production ML feature rollout