Each semester, the Dan E. Wells Outstanding Dissertation Award is presented to a doctoral student at the University of Houston College of Natural Sciences and Mathematics whose work exemplifies exceptional research quality and scholarly impact. The Fall 2025 recipient is Rabimba Karanjai, a Ph.D. graduate in computer science.
The award was announced during the NSM Commencement ceremony, where Karanjai received a certificate and a $1,000 prize in recognition of his achievement.
Karanjai’s dissertation, titled “Teaching AIs to Reason and Code, Confidentially,” addresses a pressing challenge in the rapidly evolving field of artificial intelligence: enabling AI systems to generate reliable, efficient, and secure computer code while safeguarding against errors and vulnerabilities.
Under the mentorship of Weidong Shi, associate professor in the Department of Computer Science, Karanjai developed a novel multi-agent AI framework that improves the accuracy and trustworthiness of AI-generated code. His system uses a coordinated set of trained models that collaboratively outline a high-level code structure, adapt it to a specified programming language, and perform rigorous verification to identify and correct mistakes that could lead to inefficiency or security risks.
Karanjai also created DeFaaS, a decentralized blockchain-based infrastructure designed to strengthen privacy and security in AI-assisted coding environments. His innovations have drawn attention from leading researchers and major technology organizations, including Google and the Mozilla Foundation.
Karanjai now serves as a Staff Machine Learning Scientist at PayPal.