About

I'm Shibly, a PhD student and Graduate Teaching Assistant in the Department of Computer & Information Science & Engineering at the University of Florida, advised by Prof. Sumit Kumar Jha.

My research asks how the reasoning of a large language model can be checked rather than trusted. I'm interested in whether a model's stated chain of reasoning is really the computation that produced its answer, and in building systems whose intermediate steps are small and legible enough for a person to read — treating faithfulness and explainability as things to be measured, not asserted. Before starting my PhD I spent several years building applied LLM and RAG systems in industry.