Photo Grader

This is a consulting project that I worked on during my fellowship at Insight Data Science. I built a tool for an online photography platform to predict album ratings before the photos are delivered to the customers. The model aims to identify signals for customer complaints that are overlooked by the photographers which could be easily improved. In this way, the auto pre-checking could enhance the overall experience of the platform for the photography buyers.

Tools used: Python, PostgreSQL, Keras, OpenCV, AWS, GCE, Docker, Flask

Web demoBlog post, Slides