Postdoc: Neural-Symbolic PDE Solvers

Los Alamos National Laboratory; Los Alamos (NM); On-site; USD 90,000 - 110,000; Full time

Benefits offered by this job

Dental and vision insurance

Free basic life and disability保险

Paid childbirth and parental leave

Learning opportunities and tuitionAss

401(k) with matching

Flexible schedules and time off

Onsite gyms and wellness programs

Relocation packages

Job summary

Los Alamos National Laboratory in Los Alamos, NM, seeks postdoctoral researchers to develop neural-symbolic methods for automatically constructing and verifying numerical solvers for time-dependent PDEs. The work spans operator splitting, conservative schemes, and neural modeling, with rigorous stability and conservation checks.

The role combines numerical analysis, scientific ML, and large-scale computing, requiring strong PDE background, programming in Python/C++/Julia/Fortran, and strong

Qualifications

• PhD in applied/computational mathematics, computer science, computational physics, engineering, or related field completed within the last 5 years or near completion.
• Demonstrated research in numerical methods for PDEs and related computational science topics.
• Strong programming skills in Python, C++, Julia, or Fortran for scientific computing.
• Experience implementing, testing, and evaluating numerical algorithms beyond coursework.
• Publications or substantial projects in numerical analysis or scientific computing.
• Ability to work independently and collaboratively in a multidisciplinary setting.

Responsibilities

• Develop and analyze global and local operator-splitting algorithms for nonlinear PDEs, including shock-dominated systems.
• Build generative proposers for symbolic solver programs using autoregressive models and tree search.
• Develop numerical and symbolic verification tools for conservation laws and stability.
• Implement reliable research software and run numerical experiments on large-scale platforms.
• Design reproducible benchmarks comparing methods with classical solvers and ML baselines.
• Publish research and collaborate with LANL scientists and partner institutions.

Skills

Numerical methods

PDEs

Python

C++

Julia

Fortran

Communication skills

Independent work

Team collaboration

Scientific computing

Education

PhD in applied/computational mathematics or related

Tools

Git

MPI

CUDA

GPU programming

Job description

Los Alamos National Laboratory in Los Alamos, NM, seeks postdoctoral researchers to develop neural-symbolic methods for automatically constructing and verifying numerical solvers for time-dependent PDEs. The work spans operator splitting, conservative schemes, and neural modeling, with rigorous stability and conservation checks.

The role combines numerical analysis, scientific ML, and large-scale computing, requiring strong PDE background, programming in Python/C++/Julia/Fortran, and strong