Abir Mohammad Hadi, Ph.D.
Assistant Professor · Computer Scientist · AI Researcher
I study how intelligent systems can learn, adapt, and act reliably—connecting foundational AI research with the challenges of the real world.
I’m a computer scientist and tenure-track assistant professor at Northern Kentucky University. My work connects artificial intelligence, autonomous software systems, and machine perception, with a particular interest in how learning systems can become more efficient, adaptive, and reliable.
I earned my Ph.D. in Computer Science from South Dakota State University in 2025, where I worked in research and teaching. Before academia, I worked as an R&D Software Engineer at Nerddevs Ltd. That experience continues to shape how I connect scientific questions with practical engineering.
Academic journeyFrom how systems reason to how they learn and perceive, my research asks what it takes to build intelligence we can rely on.
How can autonomous systems interpret requirements, collaborate, and produce behavior we can meaningfully check? I explore agentic AI and autonomous software engineering, with verification as a central research question.
Intelligence should adapt to both the task and the resources available. My research connects neural network compression and adaptive model complexity with the broader pursuit of learning systems that evolve over time.
Connecting learning algorithms to physical environments through computer vision and multimodal representations. Research themes include anomaly detection, infrastructure inspection, and the interpretation of heterogeneous sensor data.
My long-term ambition is to build systems that continually learn and adapt while maintaining formally specified, verifiable guarantees about their behavior. Current work on agents, verification, continual learning, and efficiency contributes toward this scientific goal.
A scientific ambition guiding current research.Work on efficient learning, machine perception, intelligent sensing, and applied AI.Google Scholar
11 publications shown
Smart Agricultural Technology, 15 · Article 102584
Moving Toward Smart, Resilient and Sustainable Bridges · pp. 667–674
Agriculture, 16(2) · Article 197
South Dakota State University
ACM SIGAPP Applied Computing Review, 24(1) · pp. 14–23
Proceedings of the 39th ACM/SIGAPP Symposium on Applied Computing · pp. 1056–1058
Proceedings of the International Conference on Research in Adaptive and Convergent Systems · pp. 1–6
Proceedings of the Conference on Research in Adaptive and Convergent Systems · pp. 107–111
International Workshop on Frontiers of Computer Vision
Journal of Information Technology Research, 14(2) · pp. 106–124
Publisher links and citation exports are provided where verified. Full texts remain subject to publisher access.
Research themes connecting published foundations with new questions in autonomous systems, adaptive learning, and sensing.
Exploring how agents with different roles and capabilities can collaborate to generate web applications.
Turning a natural-language requirement into reliable software involves interpretation, coordination, and evaluation. This theme examines how agent specialization and collaboration could support that process.
Examining functional correctness and design diversity in software produced by AI systems.
A plausible interface is only one part of a working application. This theme considers how explicit requirements and behavioral checks can help evaluate generated systems, alongside the diversity of their designs.
Investigating the relationship between task complexity, network structure, and efficient computation.
Resource constraints make model size consequential. Published work investigates task-dependent complexity and reinforcement learning for pruning; this theme connects those questions to adaptive learning.
Exploring thermal imagery and intelligent sensing for understanding infrastructure condition.
Inspection requires interpreting observations in their physical context. This theme considers how computer vision and thermal sensing can support the analysis of infrastructure without assuming that every anomaly has the same cause.
Combining heterogeneous representations to understand anomalies in imagery.
Different representations capture different aspects of a scene. Published work on satellite imagery provides a basis for asking how complementary information can inform anomaly classification.
I want students to understand not only how to build a system, but why it works. My teaching connects clear foundations, hands-on implementation, and careful reasoning—giving students the tools to ask better questions and engineer thoughtful solutions.
Organizing data, reasoning about complexity, and choosing algorithms with care.
Building the logic, data handling, and foundations behind web applications.
Creating accessible, responsive experiences with the building blocks of the web.
Designing and implementing full-stack applications using modern technologies and framworks.
Understanding the evolution of natural language processing methods.
Computer Science and Software Engineering · School of Computing and Analytics
Doctoral education, with graduate research and teaching experience.
Research and instruction in computer science.
Research and instruction in computer science.
Software research and development.
Software research and development.
I welcome conversations about research collaboration, intelligent systems, and student research interests. If your questions intersect with mine, I’d be glad to hear from you.