Test-Time Risk Adaptation with Mixture of Agents
NeurIPS 2026
UT Austin ECE ยท Ph.D. student
AI you can trust with consequential decisions.
I build AI systems that people can trust with real decisions. My research works toward that goal in three ways: verifying safety before deployment, adapting model behavior at inference time as objectives or constraints change, and evaluating methods by the decisions they produce. I apply these ideas to reinforcement learning, language models, optimization, and power systems, and develop forecasts that improve downstream decisions.
I am a Ph.D. student in Electrical and Computer Engineering at The University of Texas at Austin, advised by Prof. Hao Zhu. I received my B.Eng. in ECE with a minor in Mathematics from the American University of Beirut, where I worked with Prof. Rabih Jabr and Prof. Sami Karaki, and I have interned at Los Alamos National Laboratory with Dr. Wenting Li, Dr. Brian Bell, and Dr. Russell Bent, and at Argonne National Laboratory with Dr. Feng Qiu and Dr. Wei Gao.
Research
Methods for learning decisions under uncertainty, distribution shift, and explicit risk specifications.
Stress testing, verification, and input-space construction for learned models used in constrained settings.
Fine-tuning and inference-time methods for language models that propose technical actions under constraints.
Power-system testbeds for studying learning, optimization, reliability, market participation, and control.
Selected Publications
NeurIPS 2026
ICML 2025
ICML 2025
Recent activity