Dr. Soheil Feizi: Building Trustworthy AI for the Future
What does it mean for artificial intelligence to be reliable, and who ensures the systems shaping our future can be trusted?
For the past 15 years, Dr. Soheil Feizi, Associate Professor of Computer Science at the University of Maryland and founder of RELAI, has been tackling these very questions. His work at the intersection of machine learning, reliability, and interpretability positions him as one of the nation’s leading experts on trustworthy AI, earning recognition not only from the scientific community but also from policymakers and industry leaders.
A Journey into AI Reliability
While earning his PhD at MIT, and later during a postdoctoral fellowship at Stanford University, Dr. Feizi explored how theoretical advances in machine learning could have real-world applications. That experience solidified his focus on research and teaching – and led him to the University of Maryland, where he now spearheads pioneering work on AI reliability.
In addition to his academic role, Dr. Feizi launched RELAI, a startup that makes reliable AI systems accessible across sectors. “We’re in an exciting time for large language models and generative AI,” he explained, “but these systems have existed in various forms for decades. What matters most now is ensuring they are dependable.”
Research at the Leading Edge
Dr. Feizi’s research focuses on developing cutting-edge tools to assess and enhance the reliability of AI systems. His recent projects explore how large language models are trained, how they respond to adversarial inputs, and what clues can be uncovered to make their reasoning more transparent.
One of these efforts, Maestro, is among the first holistic optimizers for AI agents — it can both evaluate other AI models and propose solutions when it finds weaknesses. This project exemplifies his broader goal: To design AI tools that can help build our trust in these rapidly expanding technologies.
Dr. Feizi’s high-impact research has drawn attention well beyond academia. In 2024, he took his AI insights to Capitol Hill, testifying before the House of Representatives’ Bipartisan Task Force on AI to help inform national policy.
Federal Support Fuels Innovation
Like many researchers working at the frontiers of science, Dr. Feizi emphasizes the vital role of federal funding in advancing AI reliability. His work has been supported by the National Science Foundation (NSF), the Office of Naval Research (ONR), and the Army Research Office (ARO), among others. He is also a recipient of the NSF CAREER Award, the ONR Young Investigator Award, the ARO Early Career Award, and the prestigious Presidential Early Career Award for Scientists and Engineers — the highest honor the U.S. government bestows on early-career researchers.
“These federal resources make high-risk, high-impact research possible,” he noted. “They also allow me to train students — future leaders who go on to launch companies, teach, and push the field forward.”
For Dr. Feizi, mentoring students and building a workforce fluent in reliable AI development is just as important as his technological breakthroughs.
Shaping the Future of Reliable AI
Whether through RELAI’s solutions, publications from his lab, or contributions to federal policy discussions, Dr. Feizi’s work helps shape how Americans understand and trust AI systems.
At a moment when artificial intelligence is rapidly permeating daily life, Dr. Soheil Feizi represents both the scientific ingenuity and the steady leadership that the nation needs to maintain its leadership role in this potentially game-changing technology.