Hadiza Umar Yusuf

AI & CPS Safety Researcher

PhD Candidate · University of Michigan-Dearborn

Hadiza Umar Yusuf

I am a doctoral candidate in Computer & Information Sciences at the University of Michigan-Dearborn and a Graduate Research Assistant at the AISE-CPS Lab. I am passionate about the safety and reliability of AI systems, especially for cyber-physical systems that operate in the real world where failures have real consequences.

My research interests include AI Safety and Ethics, Falsification of AI-enabled Cyber-Physical Systems, Reinforcement Learning, and Trustworthy AI. I am also interested in the intersection of AI and education, specifically in improving access to learning through adaptive AI technologies.

AI Safety Cyber-Physical Systems Reinforcement Learning Software Verification Trustworthy AI

View My Research Download CV

Research & Projects

Current — Active Research

Falsification of AI-enabled Cyber-Physical Systems

Developing a falsification technique that combines stochastic optimization and reinforcement learning to improve fault detection in AI-enabled CPS. Conducting empirical analysis to categorize CPS models and identify verification challenges specific to AI systems.

📍 University of Michigan-Dearborn 👩‍🏫 Supervisor: Prof. Khouloud Gaaloal 📅 Sept. 2024 – Present

CPS Safety & Verification

Completed · CASA Prize 2026

SafeAlert

A pre-procurement AI safety evaluation kit for language models used in Nigerian fintech and mobile money contexts. Helps teams test whether a model refuses harmful fraud-generation requests and correctly classifies Nigerian fintech messages before deployment. Winner of the Africa AI Safety Prize Competition 2026.

📅 2026 · 1st Place GitHub →

Completed · UM-D

Counterfactual-Guided Debugging of Cyber-Physical Systems (DeCaF)

Developed a framework for debugging CPS using counterfactual-guided explanations that identify minimal input changes preventing system failures. Evaluated across multiple Simulink CPS benchmarks.

👩‍🏫 Prof. Khouloud Gaaloal 📅 May – Aug. 2025 · Submitted

Completed · CMU Africa

Prosthetics Design for Low-Resource Environments

Improving access to customized prosthetics using smartphone-based digital reconstruction and data-driven prediction of mechanical functionality in resource-constrained communities.

👩‍🏫 Prof. Moise Busogi 📅 May 2023 – Aug. 2024

Completed · CMU Africa

Data-Driven Material Prediction for 3D Printing

Predicted material properties of 3D-printed parts using multi-layer perceptron (MLP) models linking specimen test results, printing conditions, and geometry. Generated and pre-processed simulation data under varying parameters.

👩‍🏫 Prof. Conrad Tucker 📅 Jan. – May 2024

Completed · CMU Africa

Adaptive Education for African Languages (ED-GPT)

Constructed an LLM pipeline for context-based answer generation for high school STEM subjects in Kinyarwanda. Achieved 98% accuracy with sentence similarity evaluation.

👩‍🏫 Prof. Azzizuddin Biyabani 📅 Oct. – Dec. 2023

Completed · CMU Africa

DTMF Demodulation using TinyML

Designed an experiment for accurate DTMF signal decoding using power spectral analysis. Captured and processed real-time signals on Arduino Nano Sense BLE and Raspberry Pi.

👩‍🏫 Prof. Mike Perkins 📅 May – Aug. 2023

Publications

SiRL: Signal-Level Reinforcement Learning for Falsifying Cyber-Physical Systems

Yusuf, H. U., Gaaloul, K., Dharmadhikari, A. · ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA ’26) · Accepted

CPSFalsificationReinforcement Learning

Towards Counterfactual Explanation and Assertion Inference for CPS Debugging

Ghazal, Z., Yusuf, H. U., Gaaloul, K. · IEEE International Conference on Software Testing, Verification and Validation (ICST 2026) · 2026

CPSDebuggingCounterfactual AI

Navigating the Shift: Architectural Transformations and Emerging Verification Demands in AI-Enabled Cyber-Physical Systems

Yusuf, H. U., Gaaloul, K. · IEEE/ACM 4th International Conference on AI Engineering – Software Engineering for AI (CAIN 2025) · 2025

AI SafetyCPS VerificationSoftware Engineering

Model-Based Verification for AI-Enabled Cyber-Physical Systems Through Guided Falsification of Temporal Logic Properties

Yusuf, H. U. · IEEE/ACM 4th International Conference on AI Engineering – Software Engineering for AI (CAIN 2025) · 2025

FalsificationTemporal LogicCPS

Performance Assessment of Two Predesigned Programmable Integrated Circuits for Efficient Bioelectricity Harvesting from Soil Electroactive Microbes

Simeon, M. I., Imoize, A. L., Yusuf, H. U. · 2025 IEEE Connecting the Unconnected Europe, Middle East, and Africa Regional Summit (CTUS-EMEA) · 2025

BioelectricityEmbedded SystemsEnergy Harvesting

Accessible Digital Reconstruction and Mechanical Prediction of 3D-Printed Prosthetics

Lee, J., Nkama, C., Yusuf, H. U., et al. · Journal of Mechanical Design, 147(6), 062001 · 2025

Computer VisionProstheticsMachine Learning

Increasing Accessibility of 3D-Printed Customized Prosthetics in Resource-Constrained Communities

Lee, J., Nkama, C., Yusuf, H. U., et al. · ASME International Design Engineering Technical Conferences (IDETC/CIE 2024) · 2024

3D PrintingAccessibilityProsthetics

Inverse-Prediction of Material Property in Fused Filament Fabrication / Fused Deposition Modeling

Lee, J., Yusuf, H. U., Tucker, C. · ASME International Design Engineering Technical Conferences (IDETC/CIE 2024) · 2024

3D PrintingMaterial ScienceDeep Learning

Towards Reliable DTMF Recognition: A TinyML Approach

Yusuf, H., Perkins, M., Nkama, C. · International Congress on Information and Communication Technology (ICICT 2024) · 2024

TinyMLSignal ProcessingEdge AI

Teaching

📐

Applied Stochastic Processes

Carnegie Mellon University Africa · Fall 2023

Graduate Teaching Assistant for 53 graduate students. Held weekly recitations, graded homework, and provided academic support.

📊

Mathematics & Statistics

Federal University of Technology Minna · 2021–2022

e-Tutor for 100 and 200-level computer science students. Average student rating: 4/5. Synchronous and asynchronous support.

💻

Programming & Web Design

Merrion IT (New Horizons College) · 2020–2022

Instructed 70+ high school students in Python, Blender, and web design. Guided hands-on project-based learning.

🔢

BASIC Programming & Numerical Analysis

College of Education Katsina-Ala · 2019–2020

Assisted in teaching, evaluating, and mentoring Computer Science students in BASIC, Microsoft Packages, and Numerical Analysis.

🔬

Science (Chemistry, Basic Science & Technology)

Talented Heroes Academy · 2013

Taught high school STEM subjects. Average student rating: 4.3/5.

🤖

AI Systems Design (EdgeML Labs)

Carnegie Mellon University Africa · Summer 2023

Developed and enhanced labs for the AI Systems Design course (EdgeML piece) as a graduate intern.

Talks & Presentations

June
2026

Keynote Speaker

AI used in systems that require constant decision-making

Encode Africa, AI Safety Action Weeks funded by Bluedot Impact · Event Page

March
2026

Session Speaker

Emerging Trends in Verification of AI-enabled Cyber-Physical Systems

International Conference on Machine Learning, Artificial Intelligence and Data Science · Speaker Profile

Aug.
2022

Presenter

Searching for and Choosing Graduate Schools

Ignite Research Academy Webinar · Watch Video

News

Aug 2026

🏆 SafeAlert won 1st Place at the Africa AI Safety Prize Competition 2026, out of over 100 submissions — a prize competition focused on building practical safeguards for beneficial AI deployment across Africa.

Jul 2026

🎉 Poster accepted at ISSTA 2026 (ACM SIGSOFT International Symposium on Software Testing and Analysis).

Feb 2026

🎉 Paper accepted at ICST 2026 (IEEE International Conference on Software Testing, Verification and Validation) — Towards Counterfactual Explanation and Assertion Inference for CPS Debugging (DeCaF), a framework that identifies minimal input changes to prevent failures in cyber-physical systems.

Jan 2026

🎓 Advanced to PhD Candidate status in Computer & Information Sciences at the University of Michigan-Dearborn.

Nov 2025

✅ Passed my PhD qualifying exams in Computer & Information Sciences at the University of Michigan-Dearborn.

Oct 2025

🤝 Served as a Student Volunteer at the 28th ACM/IEEE International Conference on Model Driven Engineering Languages and Systems (MODELS 2025) in Grand Rapids, MI — one of the premier venues for model-driven engineering research.

Sep 2025

🎤 Invited as Session Speaker at the International Conference on Machine Learning, Artificial Intelligence and Data Science — to present on Emerging Trends in Verification of AI-enabled Cyber-Physical Systems.

Aug 2025

🙌 Joined the LifeWise Academy Dearborn as a Leadership Board Member, contributing to the mission of equipping students with character, wisdom, and a strong academic foundation.

Apr 2025

🎓 Joined the Wolverine Mentors Collective (Graduate Students Edition) as a mentor, supporting early-career graduate students in navigating research, academia, and professional development at the University of Michigan.

Sep 2024

🎉 Began my PhD in Computer & Information Sciences at the University of Michigan-Dearborn, joining the AISE-CPS Lab under the supervision of Professor Khouloud Gaaloul. My doctoral research focuses on falsification and verification of AI-enabled Cyber-Physical Systems — funded by the National Science Foundation (NSF).

May 2024

📰 Featured in a Mastercard Foundation Impact Story: "Everyone has something to give in their own little space" — a story about my journey through academia, research, and giving back to communities across Africa.

May 2024

🏆 Graduated from Carnegie Mellon University Africa with an MS in Engineering Artificial Intelligence, receiving two prestigious awards: the Lakshimi Subramanian CMU-Africa Student Excellence Award — recognising outstanding academic and research contributions — and the Mastercard Foundation Scholars Impact Award, celebrating scholars who demonstrate exceptional commitment to transforming Africa through their work and leadership.

Apr 2024

🎊 Received admission to the PhD program in Computer & Information Sciences at the University of Michigan-Dearborn — set to join the AISE-CPS Lab to work with Professor Khouloud Gaaloul on NSF-funded research in AI safety and cyber-physical systems verification.

I write The AI Safety Brief; a LinkedIn weekly newsletter on AI safety, alignment, and ethics. The letter discusses different AI safety research papers, articles and incidences, to spur curiousity and open up ideas for those seeking research direction in AI safety.

Subscribe on LinkedIn →

🛡️

The AI Safety Brief

When You Lock the Gate but Leave the Fence Broken

When AI safety alignment fails not through a direct attack, but through something as routine as a writing task — what does that tell us about how robust our safety tools really are?

⚙️

The AI Safety Brief

The Model That Could Not Be Released

Anthropic built a model capable enough to find decade-old security vulnerabilities on its own — and then decided not to release it. A closer look at what that decision reveals about where AI safety stands today.

🤝

The AI Safety Brief

Trustworthy AI is not just a technical problem

Building AI systems people can trust involves more than good algorithms — it requires ethics, transparency, and accountability.

Hadiza Umar Yusuf

Research & Projects

Publications

Teaching

Talks & Presentations

News

Newsletter

Get in Touch