We are recruiting two postdoctoral researchers (2 × 2-year appointments) at Eindhoven University of Technology (TU/e) within the Horizon Europe project SIMU-LINGUA, which develops scientific foundation models (SciFMs) for materials science. The project aims to advance AI for scientific discovery by combining large-scale scientific data, knowledge graphs, machine learning, and GPU/HPC computing. The positions are organized around two complementary research tracks:

Track A — Scientific Data & Knowledge Graphs
Development of scalable scientific data infrastructures, including materials ontologies, data ingestion and curation pipelines, and multi-modal knowledge graphs for training foundation models.

Track B — GPU-Accelerated Scientific Foundation Models
Design and training of large-scale machine learning models, including SciFM architectures, PyTorch GPU training pipelines, and distributed training on HPC systems.

We welcome applicants with backgrounds in machine learning, data science, applied mathematics, scientific computing, computer science, or related fields, and experience with areas such as deep learning, scientific datasets, knowledge graphs, or GPU/distributed computing.