Gen AI Powered Conversational BI

Led the development of a conversational BI tool that enabled stakeholders to interact with payments data in Amazon Redshift using natural language. The solution, inspired by ChatGPT-style interfaces, significantly improved data accessibility for non-technical users, leading to faster and more informed decision-making. Key Achievements: Designed and deployed a Streamlit-based application on Amazon ECS with a load balancer, ensuring scalability and reliability. Implemented an LLM-powered agentic workflow that translated user queries into SQL, retrieved relevant data, applied business logic, and delivered actionable insights. Conducted prompt engineering to fine-tune query responses for domain-specific accuracy. Empowered stakeholders to explore their payments data autonomously, reducing dependency on data teams and accelerating business insights. Contributed to improved decision-making and operational efficiency—reduced stakeholder reporting turnaround time by ~40%.