cv

Ph.D. in Engineering (Computer Engineering)

• MS-leading-to-PhD program, Electrical and Computer Engineering Department. CGPA 4.14/4.5.
• Thesis: Accessible and Scalable Visual-Art Appreciation Methods for Visually Impaired People.
• Coursework: Image Processing, Neural Networks, Machine Learning, AI-Based Simulation, Advanced AI, HCI Design, Estimation Theory, Advanced Computer Networks, Advanced Digital Communications.

B.E. in Electronics Engineering

• CGPA 3.59/4.
• Final Year Project: Disabled Patients Facilitation System.

Post-Doctoral Fellow, SDAIA-KFUPM Joint Research Center for AI

• Conducting research in computer vision and medical imaging, focusing on inverse problems and efficient deep learning, including diffusion-based models for image reconstruction, enhancement, and downstream vision tasks.
• Investigating resource-efficient and scalable methodologies for applying deep generative models in real-world and medical imaging applications.
• Co-supervising MS and PhD students in AI research in collaboration with faculty.
• Teaching undergraduate computing courses and delivering hands-on training in Data Science, Machine Learning, and Deep Learning, including workshops and intensive crash courses.
• Contributing to the organization of research activities and academic events within the SDAIA-KFUPM AI research center.

Senior AI Engineer, AI Engineering Division

• Developed and optimized AI-based digital radiography solutions for medical imaging applications.
• Developed weakly- and semi-supervised deep learning algorithms for triage of abnormalities and pathological findings.

Researcher, H-Lab

• Developed LSTM-based pedestrian dead reckoning for indoor positioning, achieving a state-of-the-art 0.78 m mean error without external infrastructure.
• Designed a sensor fusion and fingerprinting framework improving localization accuracy by 87%, plus a BLE RSSI-based localization technique with 89% accuracy.
• Proposed an STFT+CNN-based islanding detection method for multi-DG systems achieving 98.76% accuracy.
• Developed a multi-sensory reconfigurable assistive platform achieving 99.17% identification accuracy, validated through user studies (SUS, NASA-TLX).

Researcher, Media System Lab

• Developed a fine-grained, holistic textual-visual multi-modal representation learning method using foundation model transformers, improving mean retrieval ranking by 47% over a CLIP zero-shot baseline.
• Developed neural attention models for automating expert annotation and ranking tasks in multi-modal information retrieval.

Teaching Assistant

• Courses: Electrical and Electronics Programming, Basics of Programming in C.
• Taught and supervised practical programming sessions; graded assignments and exams.

Engineer, Multimedia and Broadband Operations

• Worked on broadband services and access network optimization and rehabilitation projects.
• Conducted MSAG, NGN, DSLAM, and ONU surveys, development, operation, and maintenance.
• Liaised with Huawei, ZTE, and Alcatel representatives for network expansion and system-level issue resolution.

Physics-Informed Data Preparation and Efficient Diffusion Modelling for High-Fidelity Medical Imaging

• Developing physics-guided diffusion architectures incorporating imaging operators (e.g., PSF, attenuation, noise models) for improved reconstruction fidelity and robustness.
• Designing computationally efficient diffusion models using few-step sampling, lightweight backbones, knowledge distillation, and quantization for real-time and edge deployment.

Spatio-Temporal Deep Learning for Robust and Efficient Face Presentation Attack Detection

• Developed spatio-temporal deep learning frameworks with squeeze-and-excitation and Bahdanau attention mechanisms for robust face anti-spoofing under real-world conditions.
• Proposed multi-modal architectures augmenting RGB with synthetic depth modalities, achieving state-of-the-art performance with near-zero error rates on benchmark datasets.

Provoking LLMs by LLMs: A Red Team Framework for Proactive Detection of Harmful Content

• Investigated red teaming strategies for LLMs to proactively identify vulnerabilities and harmful behaviors, and contributed to evaluation frameworks for robustness, safety, and trustworthiness.

Development of Non-Visual Information Transmission Interfaces for the Blind and Visually Impaired

• Developed a reconfigurable tactile platform as an assistive device for non-visual information transmission.
• Developed the AI-Ekphrasis framework using transformer-based foundation models for multi-modal image and fine-grained text retrieval.

Development of Robustness and Precision Improvement Technologies in Complex Indoor Location Estimation

• Developed and tested machine and deep learning-based solutions for robust indoor positioning and navigation, including an LSTM-based heading distance/direction estimator with 0.782 m mean error.

Best Paper Award

Best Paper Award

Best Paper Award