The Center for Hierarchical and Robust Modeling of Non-Equilibrium Transport (CHaRMNET), housed in the Department of Computational Mathematics, Science and Engineering (CMSE) at Michigan State University, seeks to hire a postdoctoral research associate in the broad area of computational physics (https://charmnet-mmicc.github.io). 

This is a one-year, benefits-eligible position that is potentially renewable on an annual basis contingent upon satisfactory performance and availability of funding.

The mission of CHaRMNET is to develop novel numerical methods that allow us to overcome the curse of dimensionality in the simulation of fusion energy systems. The curse of dimensionality refers to the fact that traditional methods scale as O(Nd), where N is the number of degrees of freedom per dimension d. Foundational kinetic models often have N∼256 and d≥6, making direct numerical simulation intractable with traditional approaches. Through innovative work combining machine learning with new paradigms for direct solvers of high-dimensional partial differential equations, members of CHaRMNET aim to overcome this challenge.

The focus of this position is on structure-preserving, machine-learning–accelerated scientific computing for plasma physics applications. In particular, the project involves developing data-driven collisional kinetic models and numerical schemes that preserve physical structures such as conservation laws and entropy dissipation. The target application is enabling scale-bridging kinetic models for fusion systems. This work encompasses both research in numerical methods for partial differential equations and machine learning methods for physics-based modeling.

Completion of a Ph.D. in mathematics, physics, or a related area is required before Fall 2026. A successful candidate will have experience with multiscale modeling, high-order methods, or fast solvers and their application to multidimensional PDEs. Relevant expertise includes, but is not limited to, kinetic equations, statistical physics, Discontinuous Galerkin methods, weighted essentially non-oscillatory methods, acceleration methods such as fast convolution, preconditioning of implicit solvers or multigrid techniques, and low-rank tensor methods for high-dimensional PDEs. A strong background in numerical analysis and multiscale modeling is essential. 

For more information, please contact Professor Andrew Christlieb if you wish to know more about this posting (christli@msu.edu).  Review of applications will begin on December 15, 2025. 

All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, citizenship, age, disability or protected veteran status.