I am a Mechanical Engineer, currently pursuing Doctoral degree at University of Connecticut. I am deeply passionate about tackling complex multi-scale and multi-physics problems through a combination of advanced computational and experimental methods. My expertise spans a wide array of advanced numerical techniques, including Finite Element Analysis (FEA), Computational Fluid Dynamics (CFD), Thermal Analysis, and Fluid-Structure Interaction (FSI). Additionally, my skills in Machine Learning, Deep Learning, and Computer Vision enable me to develop state-of-the-art solutions for challenging engineering problems.
My research interests are diverse and encompass computational geometry, optimization, and data-driven models, including Physics-Informed Neural Networks (PINNs), Neural Operators, and Bayesian Neural Networks. I am also deeply engaged in Parallel Computing with GPU technology and sophisticated randomized algorithms. My focus is on optimizing design processes and manufacturing techniques, and improving failure prediction through these advanced computational methods. Additionally, I explore probabilistic computational mechanics and multi-scale uncertainty quantification to address the complexities of modern engineering challenges. As a dedicated problem solver, I am committed to collaborating with diverse teams to deliver impactful results and push the boundaries of engineering forward.
Simultaneous Packaging and Routing of Complex Interconnected Systems. Independently, each problem is known to be NP-hard. Solving the combined problem is a particularly interesting and challenging task.
Proximity queries in dynamic environments are difficult due to constant changes in spatial data. We exploring the possibilities of novel data structure enhances efficiency in managing these queries, enabling better real-time decision-making.
Our research focuses on developing an attention-aware physics-guided neural network designed to predict fatigue in additively manufactured components. By integrating fundamental physical principles with advanced image modalities, we aim to significantly enhance predictive accuracy and improve interpretability of the model's predictions.
We are investigating parallelism in both CPU and GPU architectures, focusing on hierarchical computational structures that often struggle with parallelism. This research aims to identify strategies for maximizing parallel efficiency while managing complex data relationships. By leveraging the strengths of each platform, we seek to enhance performance in intricate computations, paving the way for faster processing in dynamic environments.