Content recommendation solution for content publishers. Making it easier for publishers to reach relevant audiences and help users to discover the content of their liking.
- Landing page characterization
- Content category classification
- Automate policy enforcement, ensuring all published campaigns adhere to the company’s policies.
Over 2 Million historical campaigns. The campaigns data include 3 components:
- Textual Content
Each component must adhere to policy, while all the components must be consistent with each other.
Python and R programming languages, Deep Learning models implemented in Tensorflow and external API services. Various NLP and Computer Vision techniques and modules were used throughout the project.
Business Impact and Insights
Production ML model automatically classifies the content category, characterizes the landing page and outputs a safety score corresponding to policy regulations. The model processed over 250K campaigns.