Visiting Research Student, MBZUAI Trustworthy Machine Learning Certified Robustness for LLM Agents Non-Convex Optimization

I am currently a visiting research student at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Abu Dhabi, working with Prof. Mohsen Guizani. My current research focuses on Trustworthy Machine Learning, including certified robustness for LLM agents, agentic AI systems, and related questions on reliability, safety, and robustness in modern AI.

I have recently submitted my Ph.D. thesis, titled "A Constrained PL-Type Framework for Non-Convex Optimization in Distributed and Online Learning", at Indian Institute of Technology Dharwad under the supervision of Dr. Bharath B.N., Associate Professor at IIT Dharwad. My Ph.D. research focused on federated learning, online learning, generalization, and optimization, with particular emphasis on non-convex optimization landscapes, distributed learning settings, and rigorous performance guarantees for modern learning algorithms.

My academic training in mathematics, especially probability and optimization, shapes my approach to machine learning research: rigorous in theory, but always attentive to computational and practical challenges. I am actively seeking postdoctoral opportunities where I can continue to deepen my research in theoretical machine learning and trustworthy AI while collaborating with interdisciplinary teams.

For more information about my work, please see my resume.

Selected Research Directions

Trustworthy Machine Learning for Agentic AI

Studying reliability, safety, and certified robustness for LLM agents and agentic AI systems.

Non-Convex Optimization in Distributed Learning

Developing PL-type frameworks for convergence and generalization in federated and decentralized learning.

Online Learning and Dynamic Regret

Analyzing learning algorithms for time-varying, non-convex environments with rigorous regret guarantees.

News

Latest Visiting research student at MBZUAI, Abu Dhabi, working with Prof. Mohsen Guizani on Trustworthy Machine Learning and certified robustness for LLM agents.