Introducing The 2021 Fellowship…

We have an exciting new group of Peter O'Donnell, Jr. Postdoctoral Fellows this year with a variety of skillsets and a shared passion for computational science and engineering. Read what they have to say about joining our interdisciplinary community.

Nicole Aretz Expertise: Model order reduction methods focused on reduced basis approximation.

I studied mathematics at RWTH Aachen University, in Germany. After writing my master’s thesis about a reduced basis approximation for 3D-VAR data assimilation, I joined the Aachen Institute for Computational Engineering Science (AICES) as a PhD student in spring 2018 to work on sensor selection methods for uncertainty quantification. I soon joined the freshly launched International Research Training Group “Modern Inverse Problems” (IRTG MIP), which is a collaboration between AICES and the Oden Institute. Motivated by my co-supervisor, Dr. Tan Bui-Thanh, I focused on Bayesian inversion to expand our sensor selection results into the probabilistic setting. Currently, I am finishing my PhD thesis and look forward to starting my fellowship at the Oden Institute.

As part of my training at IRTG MIP, I worked with Dr. Bui-Thanh’s group at the Oden Institute for 6 months in 2019. I was impressed by the strong, multidisciplinary research output of the Institute and the frequent exchange among researchers of various fields. I’m excited to join this environment again to contribute myself and connect with so many interesting people.

I will be working with Dr. Karen Willcox, whom I first met when I organized her talk at the Charlemagne Distinguished Lecture Series (CDLS) at AICES. In discussions with our student groups, I was very impressed at how quickly she would understand our topics and link them to other areas of her expertise. Over the years, Willcox’s group has continuously explored new ideas and achieved a great diversity of methods in different fields of computational engineering. As part of my postdoc position, we will employ model order reduction methods to efficiently incorporate physical dynamics into a digital twin framework and thereby improve the twin’s predictiveness and robustness against measurement noise.

Long-term, I would like to pursue a career in academia working on the mathematical and physical foundation of computational methods. The postdoc position at the Oden Institute is an excellent opportunity towards achieving this goal.

Georgios Bakirtzis Expertise: Cyber-physical systems design and analysis.

Prior to the Oden Institute I was a graduate student in the Link Lab, a cyber-physical systems research center at the University of Virginia.

I have a PhD in computer engineering. Now, I have come to the Oden Institute because it offers a lot of opportunities for me to grow and pivot my research to something adjacent to what I did in my doctorate while simultaneously branching out and learning completely new things. I am excited to learn from the diverse voices within the Institute and be in an environment where interdisciplinary research seems to be the norm rather than the exception. I believe I thrive best in such conditions.

I have a number of key goals as part of my research. I want to work on certification of sociotechnical systems with learning in the loop such that we know that the systems we deploy are assured to be safe and they will not cause us harm.

In addition, I’m studying compositional methods in reinforcement learning so we can better understand what those systems should and shouldn’t do, and multi-agent mission modeling with a more constructive approach to proof theory in this domain.

Long term, I would really love to stay in the academic space. I find teaching and research the only time I feel like time flows as if I don't exist, completely entrenched in thinking, which is very appealing to me.

Hsi-Wei Hsieh Expertise: Shape analysis, computational anatomy, computer vision, optimal control theory and machine learning.

Prior to joining the Oden Institute, I was a PhD student in Applied Mathematics and Statistics at Johns Hopkins University.

The Oden Institute provides a very nice interdisciplinary research environment for applied math researchers like myself. I'll be working with Dr. Yen-Hsi Richard Tsai on vision-based sensor placement and robotic navigation problems and related problems in computer vision, deep learning, and optimal control.

These are all interesting problems that combine knowledge from various areas, and yet there is great potential for the algorithms and overall mathematical theory to be further explored and developed. I look forward to continuing my career in applied math research in academia or industry upon completion of my postdoc.

Stefan Henneking Expertise: High Performance Computing and Finite Element Methods.

Before becoming a Peter O'Donnell, Jr. Postdoctoral Fellow, I was a PhD student in the Oden Institute's CSEM program. Under the supervision of Dr. Leszek Demkowicz, I developed a mathematical model and implementation for the simulation of fiber laser amplifiers. Prior to that, I completed a MSc in Computational Science and Engineering at Georgia Tech.

When I joined the Institute in 2016, I came primarily for the Institute's unique multidisciplinary CSEM PhD program that has a strong focus on applied mathematics. I decided to apply for a postdoctoral position at the Oden Institute because, with its prolific collaborative research community and excellent administrative staff, the Oden Institute provides the ideal environment for cutting-edge research in computational science.

As a postdoc, I aim to build a computational framework for predictive tsunami simulation. In this effort, Bayesian inversion methods and stochastic optimization, including quantification of model and parameter uncertainty, play a key role. Dr. Omar Ghattas has been a leader in these research fields; his group at the Center for Computational Geosciences and Optimization is at the forefront of developing scalable algorithms for challenging geosciences applications.

I am passionate about tackling difficult research problems in computational science with a high impact on society. In the near future, I hope to leverage and broaden my skillset to push the frontiers of mathematics and algorithms applicable to geoscientific models and beyond. I can envision a career in industry, academia, or a national lab.

Haiyi Wu Expertise: Nanofluidics and the fundamental physics of interfacial phenomena using physics-based simulations and machine learning.

I got my PhD in mechanical engineering from Virginia Tech in May 2020. After that, I worked as a Postdoc Research Associate at the Bechman Institute at the University of Illinois at Urbana Champaign.

The Oden Institute has a solid research background for developing outstanding interdisciplinary research work. Pursuing postdoctoral training here will prepare me well for an academic career in the future.

I will work with Dr. Narayana Aluru in developing a deep learning-based computational algorithm to relate the atomic force maps from molecular dynamics (MD) simulation to the force fields for molecular modeling. The research topics already being addressed in Dr. Aluru’s group are both relevant and very interesting, and Dr. Aluru has a very solid research background in computational nanotechnology and machine learning.

Milinda Shayamal Fernando Expertise: Algorithms and computational methods for next-generational supercomputers.

Before joining the Oden Institute, I was a graduate student pursuing my PhD at the School of Computing, University of Utah.

The Oden institute is a great fit for the interdisciplinary nature of my research and for collaborating with researchers in different disciplines. Besides being one of the leading institutes for computational science and engineering research in the world, the Institute also provides opportunities to help me succeed in my research career.

New discoveries in science and engineering are primarily driven by computer simulations (in lieu of physical experiments). In many cases, such as Gravitational Wave (GW) astronomy, physical experiments are impossible. While computing resources have grown exponentially in the modern computational era, they have become increasingly complex with ever-increasing heterogeneity and fine-grain parallelism, making their use by domain scientists increasingly difficult.

The key objectives of my research are ease of use by domain scientists (by using symbolical interfaces and automatic code generation), portability (ability to use across different architectures), performance (efficient use of computing resources), and scalability (ability to solve larger problems on next-generation machines). In my research, the main driving application has been computational relativity and GW astronomy. Still, my research contributions are fundamental and have also significantly impacted other areas, such as Computational Fluid Dynamics (CFD).

With Dr. George Biros, I am working on developing scalable algorithms and computational methods for solving the Boltzmann equation, which can be challenging due to higher dimensionality and other numerical constraints. The above will contribute to the computational techniques for plasma torch modeling.

With the help of Dr. Omar Ghattas, I am continuing my research work on computational relativity, explicitly focusing on inverse problems in computational relativity and approximation models for gravitational wave modeling.

I am passionate about learning, teaching, conducting research, and applying that research to problem-solving and understanding the world around us. I want to continue working in academia, improving my skills as a researcher and a mentor.