Postdoctoral Scholar

Berkeley Lab's Center for Advanced Mathematics for Energy Research Applications (CAMERA) has a new opening for a postdoctoral scholar to develop cutting-edge mathematics and algorithms to analyze complex data from Department of Energy (DOE) experimental facilities.

This role involves research and development spanning areas such as optimization, Fourier analysis, numerical linear algebra, statistics, machine learning, and high-performance computing for one or more of the following: (1) reconstruction of 3D+ structure, heterogeneity, and/or dynamics from scattering and/or microscopy data; (2) autonomous analysis and decision making for self-driving and/or human-in-the-loop experiments; (3) computer vision for extracting complex patterns, structure, and meaning from images and/or volumes; and (4) new mathematics and algorithms leading to new applications of machine learning and artificial intelligence to analysis of experimental data. Of particular interest will be new approaches for tackling multimodal data, quantifying uncertainty, providing rigorous theoretical guarantees, and modelling complex physics, noise processes, and measurement error. You will work closely with mathematicians, software engineers, physicists, materials scientists, and beamline scientists to implement these new tools on HPC computer architectures and deliver them as user-friendly software to meet DOE experimental facility needs.

You will:

• Conduct independent and collaborative research to develop new mathematics and algorithms for analyzing complex data from DOE experimental facilities.
• Develop new mathematical algorithms targeting one or more focus areas: (1) 3D+ reconstruction of structure/heterogeneity/dynamics from scattering and/or microscopy data; (2) autonomous analysis and decision-making for self-driving and/or human-in-the-loop experiments; (3) computer vision for extracting patterns, structure, and meaning from images and/or volumes; and (4) new mathematics and algorithms that enable new reliable new applications of machine learning and artificial intelligence to experimental data analysis.
• Make advances in one or more of: multimodal data fusion/joint inference; uncertainty quantification with realistic noise/measurement error; complex physics- and artifact-aware forward modeling; and theory-grounded guarantees for proposed algorithms.
• Collaborate with scientific users and experimentalists at DOE experimental facilities to apply the developed software to real datasets and meet their scientific needs.
• Publish results in peer-reviewed venues, present at conferences/workshops, and contribute to CAMERA's collaborative research activities.

We are looking for:

• Ph.D. in Applied Mathematics, Computer Science, Physics, or related field.
• Strong research track record developing advanced mathematical and computational methods for analyzing complex experimental or imaging data.
• Demonstrated expertise in several of the following areas: inverse problems, statistics, optimization, uncertainty quantification, and/or computer vision/machine learning.
• Strong foundation in at least one of: numerical linear algebra, Fourier/spectral methods, scientific computing, and/or high-performance computing.
• Proven ability to publish in peer-reviewed venues and present research at seminars, workshops, and scientific conferences.
• Excellent written and verbal communication skills, with the ability to contribute effectively to large, collaborative, multidisciplinary projects in a diverse environment.

Desired skills/knowledge:

• Familiarity with modern machine learning methods and software, including experience applying them to scientific or experimental datasets.
• Experience collaborating with domain scientists to analyze real experimental data and translate scientific questions into robust, actionable computational approaches.