Research Fellow in Machine Learning for Hydroclimatology
| Posting date: | 02 February 2026 |
|---|---|
| Salary: | £36,130 to £39,355 per year |
| Hours: | Full time |
| Closing date: | 02 March 2026 |
| Location: | Southampton, Hampshire |
| Remote working: | On-site only |
| Company: | University of Southampton |
| Job type: | Contract |
| Job reference: | 3161625WR-R |
Summary
We are seeking a highly motivated Research Fellow to join the Hydroclimatology Group at the University of Southampton, led by Professor Justin Sheffield.
You will lead the development and application of innovative Machine Learning (ML) frameworks to understand and predict the global hydrological cycle. The role will require bridging the gap between process-based physical modeling and scalable deep learning. You will contribute to several high-impact projects addressing hydrological extremes and their feedbacks within climate and human systems. Your work will have real-world impact, providing decision support for water-energy-food-health problems and enhancing early warning systems for global hazard risks.
The role will consist of:
Methodological innovation: Develop cutting-edge ML models, including hybrid physics-informed approaches, to improve the estimation, monitoring and prediction of hydrological variables.
Big data integration: Process and fuse multi-terabyte datasets, including satellite products, in-situ observations, and climate model ensembles.
Process understanding: Develop and test hypotheses about the variability and interactions of hydrological processes across scales.
Collaborative research: Work alongside national and international stakeholders to translate methodological innovations into understanding and tools that improve water resource management, sustainable development of water-energy-food systems, hazard early warning, and reduction of health and environmental impacts.
Dissemination: Lead the preparation of high-impact manuscripts for peer-reviewed journals and present findings at international conferences.
Required qualifications and experience: To succeed, you will hold a PhD (or be close to completion) in Computer Science, Applied Mathematics/Statistics, Physics, Meteorology, Hydrology, or a related quantitative field. You will have technical expertise in one or more of the following:
Advanced ML: Expertise in Deep Learning architectures, particularly those suited for spatiotemporal data (e.g., CNNs, LSTMs, Transformers, or Graph Neural Networks).
Hybrid modeling: Experience with physics-informed machine learning or the integration of ML with data assimilation/multivariate statistics.
Software frameworks: Excellent programming skills in Python, R or similar, with experience in frameworks such as PyTorch, TensorFlow, JAX, etc.
HPC & Big Data: Proficiency in high-performance computing (HPC) environments and experience with geospatial libraries (e.g., Xarray, Dask).
Desired Experience:
Prior experience with quantifying and understanding climate variability and extremes (floods, droughts, heatwaves, …).
Knowledge of uncertainty quantification and probabilistic forecasting.
Familiarity with sectors such as water resources systems, disaster risk mapping, agriculture, water-dependent energy systems, or ecosystem services.
About Us
This position is based in the School of Geography and Environmental Science. You will join a supportive, world-class research group within a University committed to fostering a culture of equality, diversity and inclusion. The School is committed to providing equal opportunities for all and offer a range of family friendly policies, flexi-time and flexible working. We are a Disability Confident employer and the School holds a bronze Athena SWAN award.
Further Information
Term: Full-time fixed term until 28 June 2028 (with potential for extension subject to funding).
Informal Enquiries: Please contact Prof. Justin Sheffield at justin.sheffield@soton.ac.uk.
We are committed to equality, diversity and inclusion and welcome applicants who support our mission of inclusivity.
Apply by 11.59 pm GMT on the closing date. For assistance contact Recruitment on +44(0)2380 592750 or recruitment@soton.ac.uk quoting the job number.
You will lead the development and application of innovative Machine Learning (ML) frameworks to understand and predict the global hydrological cycle. The role will require bridging the gap between process-based physical modeling and scalable deep learning. You will contribute to several high-impact projects addressing hydrological extremes and their feedbacks within climate and human systems. Your work will have real-world impact, providing decision support for water-energy-food-health problems and enhancing early warning systems for global hazard risks.
The role will consist of:
Methodological innovation: Develop cutting-edge ML models, including hybrid physics-informed approaches, to improve the estimation, monitoring and prediction of hydrological variables.
Big data integration: Process and fuse multi-terabyte datasets, including satellite products, in-situ observations, and climate model ensembles.
Process understanding: Develop and test hypotheses about the variability and interactions of hydrological processes across scales.
Collaborative research: Work alongside national and international stakeholders to translate methodological innovations into understanding and tools that improve water resource management, sustainable development of water-energy-food systems, hazard early warning, and reduction of health and environmental impacts.
Dissemination: Lead the preparation of high-impact manuscripts for peer-reviewed journals and present findings at international conferences.
Required qualifications and experience: To succeed, you will hold a PhD (or be close to completion) in Computer Science, Applied Mathematics/Statistics, Physics, Meteorology, Hydrology, or a related quantitative field. You will have technical expertise in one or more of the following:
Advanced ML: Expertise in Deep Learning architectures, particularly those suited for spatiotemporal data (e.g., CNNs, LSTMs, Transformers, or Graph Neural Networks).
Hybrid modeling: Experience with physics-informed machine learning or the integration of ML with data assimilation/multivariate statistics.
Software frameworks: Excellent programming skills in Python, R or similar, with experience in frameworks such as PyTorch, TensorFlow, JAX, etc.
HPC & Big Data: Proficiency in high-performance computing (HPC) environments and experience with geospatial libraries (e.g., Xarray, Dask).
Desired Experience:
Prior experience with quantifying and understanding climate variability and extremes (floods, droughts, heatwaves, …).
Knowledge of uncertainty quantification and probabilistic forecasting.
Familiarity with sectors such as water resources systems, disaster risk mapping, agriculture, water-dependent energy systems, or ecosystem services.
About Us
This position is based in the School of Geography and Environmental Science. You will join a supportive, world-class research group within a University committed to fostering a culture of equality, diversity and inclusion. The School is committed to providing equal opportunities for all and offer a range of family friendly policies, flexi-time and flexible working. We are a Disability Confident employer and the School holds a bronze Athena SWAN award.
Further Information
Term: Full-time fixed term until 28 June 2028 (with potential for extension subject to funding).
Informal Enquiries: Please contact Prof. Justin Sheffield at justin.sheffield@soton.ac.uk.
We are committed to equality, diversity and inclusion and welcome applicants who support our mission of inclusivity.
Apply by 11.59 pm GMT on the closing date. For assistance contact Recruitment on +44(0)2380 592750 or recruitment@soton.ac.uk quoting the job number.
Proud member of the Disability Confident employer scheme
Disability Confident
About Disability Confident
A Disability Confident employer will generally offer an interview to any applicant that declares they have a disability and meets the minimum criteria for the job as defined by the employer. It is important to note that in certain recruitment situations such as high-volume, seasonal and high-peak times, the employer may wish to limit the overall numbers of interviews offered to both disabled people and non-disabled people. For more details please go to Disability Confident.