This project analyzes news article titles from major international news outlets by converting them into vector embeddings, reducing dimensionality, clustering them, and extracting representative feature words. 

This project analyzes the factors driving a customer's historical purchase value. By identifying the most influential features through SHAP analysis, the project provides data-backed insights for optimizing future marketing spend and maximizing Customer Lifetime Value.

To fine-tune the pre-trained model CardiffNLP/twitter-roberta-base-sentiment-latest in order to improve its ability to classify text sentiment into three categories: Positive, Neutral, and Negative.

[Kaggle Hackathon] This project implements a Multi-Agent Compliance System designed to mitigate the risks associated with LLM Hallucination by ensuring all legal advice is grounded in RAG-retrieved GDPR articles. 

This project allows you to create a visually stunning portfolio entirely from AI-generated images. Using Stable Diffusion, images are generated based on text prompts, saved automatically, and displayed in a grid format suitable for Markdown or HTML embedding.