Researcher (f/m/d)

Zuse-Institut Berlin

Full-time

Temporary

61,193.82 - 86,189.54 € per year

Position Details

For the Applied High-performance Computing working group of the Distributed Algorithms and Supercomputing department, we are offering a position, starting as soon as possible, on a full-time basis (39,4 hours per week), limited until September 30, 2027. If the applicant meets the relevant wage requirements and personal qualifications, the salary will be based on remuneration group 13 TV-L of the pay scale for the German public sector.

Background & Context

We are looking for a motivated early-career researcher who has completed a Master's degree in computational science, physics, or chemistry and wants to work at the interface of electronic structure theory, machine learning, and high-performance computing.

The position is embedded in an interdisciplinary team of physicists, chemists, and computer scientists working on chemical structure prediction. The work has two parts: an electronic-structure study and a supporting role in the group's dataset and machine learning activities. Both are carried out on NHR@ZIB's HPC systems.

Your Tasks

The successful candidate will conduct research at the intersection of mathematical optimization, first-order methods, and heuristic algorithm design. The work combines theoretical foundations, algorithmic development, and software implementation in an oracle-based solver framework.

Completion of a DFT/TDDFT study of molecular electronic and excited-state properties using Gaussian 16 and related quantum chemistry codes, and prepare the results for publication in a peer-reviewed journal
Performance of high-throughput reference calculations on HPC resources, including the setup and automation of workflows, submission and monitoring of jobs, and quality control of the resulting data
Management and preprocessing of large molecular datasets for model training
Support for the training, validation, and testing of neural network models, including the performance of experiments, benchmarks, and ablation studies in PyTorch as well as the analysis and documentation of results