Applied Mathematics Postdoc Adaptive Physics-Constrained Generative AI and Control

Position: Applied Mathematics Postdoc for Adaptive Physics-Constrained Generative AI and Control
** What You Will Do*
* The Instrumentation & Controls Group (AOT-IC) is seeking an Applied Mathematics Postdoctoral Research Associate in the area of adaptive, physics-constrained generative Artificial Intelligence (AI) for time-varying dynamic systems, with a focus on electrodynamics and charged particle beams. The position is part of a new project to develop generative diffusion/flow-based digital twins of complex systems that must be tracked in real time using only sparse, non-invasive measurements.

Project work includes the development of adaptive and physics-constrained generative diffusion/flow models for time-varying systems, such as the 6D phase space of an intense charged particle beam, as it drifts beyond the span of the training data, without retraining. The work ranges from theory (convergence, stability, and robustness of feedback-guided diffusion/flow processes that evolve jointly in synthetic diffusion time and in the physical time of the system being tracked) to algorithm and software development to experimental demonstration on operating accelerators.

Model development will utilize experimental data as well as synthetic data generated by electrodynamics and particle accelerator beam dynamics codes. The successful candidate will develop adaptive diffusion/flow-based virtual diagnostics and digital twins for the LANSCE accelerator, beginning with the beamline that serves the proton radiography (pRad) facility and expanding upstream to the injector and linac, and will couple them with adaptive feedback controllers running on the accelerator control system for non-invasive diagnostics and autonomous tuning, optimization, and control of intense charged particle beams.

Prior experience with particle accelerators is welcome but not required; candidates with strong backgrounds in control theory, stochastic analysis, or generative modeling are encouraged to apply.

AOT-IC regularly collaborates with particle accelerators around the world. Collaboration and experimental data will include data sets collected from intense linear accelerators at LANL such as the LANSCE linear proton and H- ion accelerator and the DARHT linear induction electron accelerator, and the European XFEL free electron laser  successful candidate will join a multidisciplinary team of accelerator physicists, applied mathematicians, computer scientists, and theorists, and will interact with researchers and application-oriented scientists on this team as well as in applied mathematics, computing, and physics groups.

Access will be provided to dedicated multi-GPU computing resources for AI development as well as state-of-the-art computing facilities within LANL.

** What You Need*
* ** Minimum

Job Requirements:

*
* + Strong applied mathematics background.

+ Demonstrated experience with deep learning frameworks (PyTorch, Tensor Flow, JAX, Keras, etc.)

+

Working experience with utilizing GPUs for machine learning.

+ PhD in related fields completed within the past five years or soon to be completed:
Applied Math, Computational Sciences, Statistics, Physics.

+ Practical experience with machine learning methods including gaussian processes and various neural network architectures (CNN, VAE, UNet, transformers, etc.)

+ Code development experience in Python.

+ Excellent communication skills (both oral and written).

** Desired

Qualifications:

*
* Graduate level knowledge of optimization and numerical methods for partial differential equations.

Experience in numerical methods and high-performance computing.

** Work Environment:*
* *
* Work Location:

** The work location for this position is onsite and located in Los Alamos, NM. All work locations are at the discretion of management.

** Note to Applicants:*
* As part of the application process, please upload your CV, a cover letter detailing work interest and qualifications, and names of three references.  
** Contact Dr. Alexander Scheinker (ascheink)
** for more details. The review process will begin immediately and continue until the position is filled. Note that US citizenship is not a requirement.

Due to federal restrictions contained in the current National Defense Authorization Act, citizens of the People's Republic of China-including the special administrative regions of Hong Kong and Macau-as well as citizens of the Islamic Republic of Iran, the Democratic People's Republic of Korea (North Korea), and the Russian Federation, who are not Lawful Permanent Residents ("green card" holders) are prohibited from accessing…