Hi, I'm
Muhammad Umar,
MS Student & ML Engineer

Advanced Machine Learning for Intelligent Systems: Development of robust, data-efficient, and generalizable deep learning models for real-world industrial and healthcare applications.

Signal Processing & Time-Frequency Intelligence: Physics-aware signal representation learning using FFT, STFT, wavelets, and adaptive transforms for machinery health monitoring.

Fault Diagnosis, Prognostics & RUL Prediction: Interpretable and domain-adaptive approaches for early fault detection, degradation modeling, and predictive maintenance.

Explainable & Trustworthy AI: Model transparency, reliability assessment, uncertainty quantification, and human-centric explainability for safety-critical systems.

Transformers, Foundation Models & Generative AI: Vision transformers, multimodal foundation models, and generative approaches for industrial inspection and biomedical imaging.

Physics-Informed & Graph-Based Learning: Integration of domain knowledge, physics constraints, and graph neural networks for smart sensing and complex system modeling.

Multimodal & Low-Label Learning: Self-supervised, few-shot, and cross-modal learning for scalable condition monitoring under limited annotated data.

Edge AI & Real-Time Intelligent Monitoring: Lightweight deep models, on-device intelligence, and efficient deployment for industrial IoT environments.

Digital Twins & Predictive Industrial Analytics: AI-enabled digital twin frameworks for system health awareness, reliability optimization, and lifecycle management.