Research Associate in AI for Environmental Forecasting (SENTINEL project)

Overview

The University of Leicester is offering a position for a Research Associate in AI for Environmental Forecasting to lead the artificial intelligence development strand of the UKRI-funded SENTINEL project. Based in the School of Engineering, the successful candidate will work on developing a physics-informed artificial intelligence forecasting system designed to monitor chemical fugitive emissions. The role involves constructing a multi-modal training dataset and designing, training, and evaluating the Receptor Transfer Function Learner. This component utilizes generative artificial intelligence downscaling, graph neural networks, and explainable artificial intelligence to predict community exposure to fugitive emissions between 24 and 72 hours in advance.

The appointee will handle real-world environmental, meteorological, operational, and community data while employing reproducible research methods with high-performance computing and graphics processing unit facilities. The project requires close collaboration with an interdisciplinary team covering artificial intelligence, atmospheric dispersion, chemical fate, and public health, alongside project partners such as the Met Office and the UK Health Security Agency. Additional responsibilities include benchmarking the system against traditional dispersion models, contributing to academic publications, and assisting in the creation of future research collaborations.

Requirements

Applicants must hold a doctoral degree, an equivalent professional qualification, or comparable experience in machine learning, computer science, environmental data science, data science, engineering, physics, or a related quantitative field. Candidates are required to demonstrate experience in developing and training deep learning or machine learning models using noisy, real-world, or time-series data, accompanied by strong analytical problem-solving abilities and a history of contributing to peer-reviewed publications or equivalent outputs. Proficiency in Python programming is essential, including familiarity with deep learning frameworks such as TensorFlow, PyTorch, or JAX, as well as scientific Python libraries like pandas, NumPy, or xarray.

Candidates must possess the capability to function independently as well as within a cooperative team, communicate intricate concepts effectively, and interact with various stakeholders. Advantageous qualifications include prior experience with generative or diffusion models, graph neural networks, physics-informed machine learning, explainable artificial intelligence, geospatial or weather data, atmospheric dispersion modeling, or utilizing graphics processing unit clusters and high-performance computing environments.

Key dates

The closing date for applications is 23 November 2026. The fixed-term contract is scheduled to begin on 01 January 2027 and run until 31 March 2028.