Postdoc: Real-Time Stokes Inversion with GPUs & ML

KIS; United States; Remote; USD 62,000 - 73,000; Full time

Benefits offered by this job

Supplementary pension

Public sector benefits

Flexible working hours

Childcare support

Public transport discounts

Christmas bonus

Job summary

Institute for Solar Physics (KIS) invites applications for a two-year postdoctoral position on fast spectro-polarimetric (Stokes) inversion techniques. The project aims to accelerate analysis with GPUs and ML, enabling near-real-time processing for several solar instruments including DKIST, VTF, and GREGOR.

Qualifications include a PhD in a relevant field and experience with radiative transfer, Stokes inversion, CUDA, and ML frameworks such as PyTorch or TensorFlow. Start date as early as Jan.

Qualifications

• Applicants should hold a PhD in physics, astrophysics, applied mathematics, or a related field.
• Experience with radiative transfer and/or Stokes inversion techniques is highly desirable.
• Familiarity with GPU programming (e.g., CUDA) and ML frameworks (e.g., PyTorch, TensorFlow) is a plus.

Responsibilities

• Develop fast spectro-polarimetric inversion techniques using GPUs and ML.
• Extend the code to instruments prioritizing GREGOR, VTF at DKIST, and related platforms.

Skills

CUDA

PyTorch

TensorFlow

Radiative transfer

Stokes inversion

GPU programming

Education

PhD in physics/astrophysics/applied mathematics

Tools

CUDA

Job description

Institute for Solar Physics (KIS) invites applications for a two-year postdoctoral position on fast spectro-polarimetric (Stokes) inversion techniques. The project aims to accelerate analysis with GPUs and ML, enabling near-real-time processing for several solar instruments including DKIST, VTF, and GREGOR.

Qualifications include a PhD in a relevant field and experience with radiative transfer, Stokes inversion, CUDA, and ML frameworks such as PyTorch or TensorFlow. Start date as early as Jan.