Interpretable machine learning for dynamic systems
System identification and model-based design
Digital twins and real-time optimization
Physics-informed and symbolic modeling approaches
Data-driven methods integrated with first-principles models
Process modeling, control, and decision support
Applications in energy, pharmaceuticals, and manufacturing
Transparent, trustworthy AI for industrial process systems
Blending chemical engineering, mathematics, and AI, my research develops methods for automatically learning interpretable differential equations and closed-form representations directly from data. By combining domain expertise with symbolic machine learning, the approaches I develop help to uncover patterns in noisy, low-data environments, expressed in ways that humans can readily understand. The resulting models contribute to transparent system identification, enhanced trust in predictions, and the creation of digital twins that are not only predictive, but also explainable. I will continue to integrate emerging AI tools with classical modeling techniques to make complex systems more understandable, robust, and accessible while requiring less effort from domain experts.