How to Apply
To apply, please email me with [phd-appl] tag in the subject line, and attach the following:
- CV
- Brief statement of interest (can be in the email body)
- Any additional materials that support your application (recommendations, research statement, etc.)
Deadline: December 15, 2024, or until the position is filled.
Please note that applications will be reviewed on a rolling basis every 2-4 weeks, so early submission is advantageous.
I reserve the right to ignore applications that do not follow the instructions, look too generic, or otherwise not serious.
After You Apply
Applications that pass the initial screening will be followed by a 1:1 interview where we will discuss your research interests and background.
Should there be mutual interest following the initial interview, you will be encouraged to formally apply and invited for a second interview. At this stage, you will also need to provide a research proposal and two letters of recommendation. The second interview will focus on evaluating your technical skills, with specific details provided via email. This interview will include at least one additional faculty member from the institute.
Fully-funded PhD in Geometric Learning and/or Uncertainty Quantification
Join our new research group at the University of Edinburgh, focusing on geometric learning and uncertainty quantification!
Our primary research aim will be to explore how the geometry of data can help build better uncertainty-quantifying models, particularly for complex data types like molecules and images. However, other research directions within geometric learning and uncertainty quantification will also be available for you to explore. This PhD position offers a unique chance to influence the trajectory of our group's work and contribute significantly to these growing fields.
This PhD position is ideal for candidates interested in the following areas of machine learning:
- Models that leverage the inherent structure of data, whether explicit or implicit [Geometric Learning]
(to improve data-efficiency and satisfy constraints) - Models that quantify uncertainty associated with predictions [Uncertainty Quantification]
(to support model-based decision-making techniques and safety-critical applications)
When: Ideally, starting September 2025.
Where: School of Informatics, the University of Edinburgh, one of the top computer science departments globally [1, 2]. More specifically, you will be a part of the Institute for Adaptive and Neural Computation, a vibrant community of researchers in machine learning and related areas.
The program lasts 3.5 years. It is straight to research, meaning no coursework. Teaching is optional. If you teach, you get additional money for it.
Funding: The position is fully-funded. Funds cover all tuition fees, and a standard UKRI-rate stipend. You will also have access to the common travel funds of the institute. Note that stipends are not taxed in the UK. Also, PhD students do not pay council tax, i.e. rent is a bit cheaper for you.
Overall, the stipend should be sufficient to live rather comfortably in Edinburgh, see, e.g., this reddit thread.
Requirements: Python programming skills, solid mathematical background. An international equivalent of the UK 2:1 honours degree (i.e. bachelor's degree with high-enough GPA) in computer science, mathematics, or a related discipline. A master's is desirable but not required. A proof of English language proficiency (can be postponed until after the offer, the offer will then be conditional on receiving an appropriate proof). See more on this page.