Postdoctoral Fellow Position
A position for Postdoctoral Research Fellow is available at the University of Michigan, Mechanical Engineering Department, UQ-SciML Group. The candidate will develop theory and algorithms for uncertainty quantification (UQ) and scientific machine learning (SciML). Research involves, for example: computational Bayes, optimal experimental design, variational inference, Markov chain Monte Carlo, information divergence estimation, reinforcement learning (RL), and human decision-making under uncertainty. These UQ/ML methods will be developed alongside applications of complex physical systems, such as fluid mechanics, material physics, and physically-based biomedicine. Thus, the candidate will work in a collaborative and interdisciplinary environment that intersects statistics, ML, computational science, and engineering physics.
Interested applicants please send your CV, representative publications, and contact information for three references to Prof. Xun Huan (email@example.com). Review of applications will begin immediately and will continue until the position is filled.
- PhD in a field of applied mathematics, statistics, computer science, or related engineering or science subject area
- Technical expertise in UQ and/or ML
- Technical expertise in at least one of: Bayesian inference, optimal experimental design, variational inference, or Monte Carlo methods
- Publication record indicative of relevant research expertise
- Excellent interpersonal, written and oral communication skills
- The use of state-of-the-art UQ/ML tools in the physical sciences
- Knowledge and expertise in Python, MATLAB, R, Julia, C, C++, or related languages
- Experience in high-performance, distributed, or parallel computing
- Knowledge and expertise in computational science and software development
- Ability to work in collaborative, interdisciplinary research environments on problems comprising diverse application domains
- A background in solving practical problems in science and engineering that involve encounters with real-world data