This repository hosts an Agentic AI Pipeline deployed on IBM watsonx designed to automate the fetching, quality-reviewing, and precision-fitting of multi-quarter Kepler light curves.
Built entirely from scratch and scaling upon proprietary methodologies developed during my undergraduate research, this system eliminates manual data cleaning and tuning. By combining a robust Python preprocessing pipeline with an LLM-driven ReAct (Reasoning and Acting) loop, the architecture ensures autonomous, end-to-end data integrity—detecting corrupted inputs and validating model convergence before finalizing downstream scientific results.