About

This project builds an automated AI Data Scientist pipeline using Zerve. The workflow ingests raw datasets, performs automated data profiling, feature engineering, and iterative model training. A self-improving machine learning loop evaluates multiple models, performs error analysis, and improves performance across iterations.

The system then segments the dataset into high, medium, and low probability groups and generates explainable insights using feature importance and SHAP drivers. Finally, an AI analysis layer converts the analytical outputs into natural-language summaries and actionable recommendations.

The goal of the project is to demonstrate how Zerve can be used to build an end-to-end automated analytics workflow that can analyze any dataset, train models, generate insights, and present results in a structured business-ready format.