Agentic AI Campaign Configuration with ReAct

This project explores the power of Agentic AI to transform natural language into structured, execution-ready configurations for enterprise-scale machine learning systems. I designed and developed an agent chatbot that enables users to describe their marketing objectives in plain English, eliminating the need to manually navigate complex campaign parameters, database metadata, and technical specifications. Built around a ReAct-based reasoning framework, the system interprets user intent, retrieves relevant information, asks clarifying questions, and generates structured job definitions through an interactive conversational interface. To ensure reliability, I also developed a dashboard for performance monitoring as well as an automated evaluation pipeline that assesses schema compliance, field-level accuracy, and hallucinations against ground-truth configurations. Achieving 94.41% overall field-level matching accuracy, the project demonstrates the feasibility of using Agentic AI to automate complex campaign configuration workflows while significantly reducing the technical expertise and manual effort required.