Research Associate (Grade 7) or Assistant Research Professor (Grade 9) in AI for Fundamental Physics (Fixed Term)

University of Cambridge; Cambridge; On-site; GBP 42,000 - 64,000; Full time

Job summary

Kavli Institute for Cosmology, University of Cambridge invites applications for a circa four-year research position in AI for fundamental physics and cosmology. The postholder will develop AI-assisted methods and GPU-accelerated inference to test theories against cosmological data, and will lead the software development required for these computations.

The role is at the intersection of theoretical and computational physics, offering substantial autonomy to shape the programme and contribute to

Qualifications

• PhD in physics, astronomy, applied mathematics or a closely related field.
• Fluency in directing AI agents and judgement to assess outputs.

Responsibilities

• Develop AI-assisted development and GPU-accelerated inference for cosmology.
• Derive physical predictions and direct the construction of the software needed to test them.
• Confront models with cosmological and astrophysical data and publish results.

Education

PhD in physics, astronomy, applied mathematics or closely related field

Tools

GPU programming

Job description

Fixed-term: The funds for this post are available until 30 September 2030 in the first instance.

Applications are invited for a circa four-year research position in AI for fundamental physics and cosmology in the Handley Lab at the Kavli Institute for Cosmology, Institute of Astronomy, University of Cambridge. Appointment will be at Research Associate level (Grade 7) or Assistant Research Professor level (Grade 9), according to the successful candidate's skills, experience and research profile. Please see the Further Particulars associated with this vacancy for more information on what is required of both the Grade 7 and the Grade 9 role.

Please note that appointment at Grade 9 is subject to application and approval to the Faculty Board. If any Faculty Board application is unsuccessful then appointment will be made at Grade 7 Research Associate.

The conventional account of fundamental physics begins with a Lagrangian, derives its observable consequences and ends with a comparison against data. In modern cosmology, the middle of this process has become the difficult part. New theories require substantial calculations, specialised numerical methods and research software capable of carrying their predictions through to cosmological observables. Much of this machinery is concentrated in large collaborations and established frameworks, while tractability at cosmological scales often relies on effective descriptions which have integrated out the physics under investigation.

This creates a bias in which theories are tested. So long as it is substantially harder to investigate a new theory from scratch than to rerun an established model, our attention will be directed towards the theories which are easiest to implement rather than those which are most scientifically promising.

The project will use AI-assisted development and GPU-accelerated inference to change what can be attempted by a small research group. The postholder will derive physical predictions, direct the construction of the software needed to test them, and confront the resulting models with cosmological and astrophysical data. The scientific judgement remains with the researcher: deciding which problems are worth pursuing, understanding what the calculations mean, and recognising when the machinery is wrong.

The position offers a larger than usual amount of research freedom. Candidates will be encouraged to shape the programme around their own interests and may come from any relevant area of theoretical or computational physics, including gravitation, field theory, lattice and numerical field theory, cosmological perturbation theory, Bayesian computation, GPU and differentiable programming, or AI-assisted scientific software development.

The postholder will also develop GPU-accelerated Bayesian inference methods, including nested sampling, publish and present their research, and contribute to the fundamental physics, cosmology and astronomy communities in Cambridge.

Applicants should have, or be close to obtaining, a PhD in physics, astronomy, applied mathematics or a closely related field. Theoretical capability is the essential criterion. Fluency in directing AI agents matters more than prior strength as a programmer, provided the candidate has the judgement to assess their outputs and correct the implementation.

Applications should include a cover letter, curriculum vitae, publication list, statement of research interests, a GitHub username (or equivalent evidence of computational skills and experience), and contact details for two academic referees. One referee should be your most recent line manager.

The covering letter should outline how you match the criteria for the post and why you are applying for the role.

Where applicants indicate that they give permission, referees will be contacted for shortlisted candidates before the interview process via the University recruitment system.

If you upload any additional documents which have not been requested, we will not be able to consider these as part of your application.

Informal enquiries are welcomed and should be directed to: Dr Will Handley Email: [e-mail]

If you have any queries regarding the application process please contact [e-mail].

The closing date for applications is: 23:59 GMT on Friday 25th September 2026.

The interview date is not yet confirmed but could be as early as week commencing 28th September 2026.

Please quote reference LG51052 on your application and in any correspondence about this vacancy.

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.

Relevant information on current research activities can be found at: https://handley-lab.co.uk https://willhandley.co.uk https://www.ast.cam.ac.uk https://www.kicc.cam.ac.uk