ML
ML Image Classifier
A machine learning pipeline for image classification using Python and TensorFlow, trained on custom datasets with 95%+ accuracy.
Context
Research project for automated quality control in manufacturing.
Role
ML Engineer
Goal
Achieve 95%+ classification accuracy on industrial components.
Overview
Built a complete ML pipeline for classifying industrial components using computer vision. The model uses a fine-tuned ResNet50 architecture trained on a custom dataset of 10,000+ images. Includes data augmentation, transfer learning, and a Flask API for inference.
Tech Stack
PythonTensorFlowOpenCVFlaskDocker
Key Features
- Transfer learning with ResNet50
- Custom data augmentation pipeline
- REST API for real-time inference
- Docker containerized deployment
- Performance monitoring dashboard