Automated Re-training of Machine Learning Models
- Developed Python scripts to automate retraining of machine learning models deployed in production without human intervention.
- Implemented a robust retraining solution that runs on a scheduled basis, executing weekly updates to ensure he models remain accurate and up-to-date with changing data patterns.
Machine Learning Prediction model using Decision Tree Classifier
- Developed accurate ML models using Decision Tree Classifier to predict GL code variables with accuracy > 80%
- Used FastAPIs to seamlessly integrate responses from ML model and other solution
Object Detection Model using YOLO
- Trained a state-of-the-art YOLO-based object detection model for invoice processing, achieving exceptional accuracy > 70% in extracting invoice fields
- Improved data extraction efficiency by accurately detecting invoice numbers, account details, and dates