User Engagement Prediction

This project focused on the design and implementation of an AI-driven system that predicts the potential performance of social media posts based on both textual and visual content before they are published.
The system enables content creators and marketers to refine their posts for maximum users engagement.
Design and implementation of a NLP-based system for predicting the performance of social media contents based on visual and textual features.
Built a predictive analytics pipeline that analyzes multimodal features, including image composition, visual appeal, sentiment, linguistic style and posting time
Applied Natural Language Processing (NLP) to evaluate text, captions, hashtags, and emotional impact
Integrated computer vision models to extract visual feature from images and videos.
Developed machine learning models to combine visual and textual features, and other metadata into a unified presentation.
Integrated and fine-tuned a prediction layer to provide probability of high/low performance.