PhD position Untangling multi-property NMR signals in drug screening with data-driven neural net...

About the role

An innovative PhD project within the Collaborative Research Center (SFB) HyPERiON at KIT, focusing on resolving signal overlap in parallel NMR spectroscopy using artificial intelligence (AI). NMR spectroscopy is a key tool in drug discovery, but signal couplings and overlaps in parallel setups complicate the extraction of critical molecular information. The project aims to develop AI models capable of generating individual, decoupled spectra from coupled NMR spectra.

This position is suitable for part-time employment. Starting date: 24.06.2026.

Responsibilities

• Develop a transformer-based neural network for processing NMR spectra.
• Create datasets from existing experiments within the CRC and from your own experiments, conducted during a research stay at KIT’s Institute of Microstructure Technology (IMT).
• Apply self-supervised pretraining based on masked sequence modeling and task-specific fine-tuning to the trained neural network.
• Analyze the extent to which the developed model can learn the underlying physical principles of nuclear magnetic resonance.
• Participate in CRC activities and engage with other PhD students and projects within HyPERiON.

Key Focus Areas

• Scalable deep learning methods for nuclear magnetic resonance.
• Self-supervised pre-training techniques and transfer learning approaches in Transformer-based architectures.
• GPU-based computing and high-performance computing (HPC).
• Application of AI methods in a scientific context.

About the Organization

This research center is part of the Helmholtz Association of German Research Centers, Germany's largest scientific organization with over 42,000 employees and an annual budget exceeding €5 billion.