Research Associate in Machine Learning for Materials Simulation (Fixed Term)

University of Cambridge; Cambridge; On-site; GBP 38,000 - 46,000; Full time

Job summary

The University of Cambridge invites a postdoctoral researcher to develop and apply machine learning methods to atomistic simulations of materials, with emphasis on amorphous and disordered systems. Based in the Theory of Condensed Matter group at the Cavendish Laboratory, under Dr Christoph Schran.

You will generate reference data, train and validate ML interatomic potentials, and run large-scale MD on national HPC facilities, with scope to influence the project and collaborate across the

Qualifications

• PhD in physics, chemistry, materials science or related discipline.
• Strong background in ML interatomic potentials: development, training and validation.
• Experience with Python and ML frameworks (e.g., PyTorch).
• Experience with amorphous or disordered materials is required.

Responsibilities

• Develop and apply ML models for amorphous materials.
• Generate reference data and validate models.
• Run large-scale MD simulations on HPC resources.
• Collaborate within Theory of Condensed Matter and across the Lennard-Jones Centre.

Skills

Machine learning

Python

PyTorch

HPC

GPU computing

Education

PhD in physics, chemistry, materials science

Tools

Density Functional Theory

Job description

Fixed-term: The funds for this post are available for 6 months in the first instance.

We are seeking a postdoctoral researcher to develop and apply machine learning methods to the atomistic simulation of materials, with a particular focus on amorphous and disordered systems, based in the Theory of Condensed Matter group at the Cavendish Laboratory working in the group of Dr Christoph Schran.

Machine learning interatomic potentials now make it possible to run molecular dynamics simulations at close to first-principles accuracy on length and time scales far beyond the reach of electronic structure methods. Amorphous materials are among the most demanding applications. Their properties depend on complex, porous structures, which places stringent demands on the models: they must remain reliable across a wide range of conditions, and they must be validated against properties that emerge only from large systems and long trajectories.

The role holder will develop and apply machine learning models for amorphous materials, covering the generation of reference data, model training and validation, and large-scale molecular dynamics simulations on national high-performance computing facilities. There is scope to shape the direction of the project and to collaborate with others in the Theory of Condensed Matter group and across the Lennard-Jones Centre.

Candidates should hold, or be close to obtaining, a PhD in physics, chemistry, materials science or a related discipline. A strong background in machine learning interatomic potentials, including their development, training and validation, is essential, as are strong scientific programming skills in Python and machine learning frameworks such as PyTorch. Direct experience of simulating amorphous or disordered materials is also required. Experience of generating reference data with density functional theory, of high-performance and GPU computing, or of contributing to open-source scientific software would be an advantage.

Appointment at research associate is dependent on having a PhD including those who have submitted but not yet received their PhD (in which case appointment will initially be made at research assistant and amended to research associate when the PhD is awarded).

Informal enquiries are welcomed and should be directed to: Dr Christoph Schran ([e-mail]). If you have any questions about the application process, please contact [e-mail].

The University actively supports equality, diversity and inclusion and encourages applications from all sections of society.

The University has a responsibility to ensure that all employees are eligible to live and work in the UK.