Fiona Fang completes her presentation at the Cambridge Mathematics Placements Presentation Day
Work on real research problems; try out your own ideas; see how a project changes as new results come in
For the fourth consecutive year, we’ve been able to offer 5 fully-funded internship positions at ICCS. Each full-time ICCS internship ran for 8 weeks during July and August, with interns paid the National Living Wage plus holiday pay.
This year, our interns were also able to attend the 5-day ICCS Summer School and benefit from the workshops and networking opportunities there. The internships concluded with talks given at the Cambridge Mathematics Placements Presentation Day at the Centre for Mathematical Sciences.
Read full details of each 2026 internship opportunity, which covered areas ranging from greenhouse gases in the Fens, ocean turbulence and symbolic execution of Fortran programmes.
Professor Colm-cille Caulfield, Co-Director (Science) at ICCS, said: “Every year it’s a privilege to host the next generation of climate researchers and software engineers here at ICCS. Now more than ever, it's so important to introduce students to major research challenges in computational climate science. Every year I am amazed by the progress our interns make during the summer and I'm excited to see what the class of 2026 does next.”
What our interns said
Fiona Fang
Project: Combining observations and machine learning to infer ocean turbulence
"I was initially drawn to this project because, from my undergraduate lectures, I had learnt that ocean turbulence remains difficult to constrain, as direct microstructure observations are sparse. I was therefore excited by the possibility of using modern machine learning approaches to help bridge this observational gap.
“During the internship, I explored two types of models, XGBoost and artificial neural networks, designing sensitivity experiments across different model configurations and predictor sets and developing metrics to evaluate their performance. My results showed that machine learning models can infer turbulent dissipation rates from hydrographic measurements, which are far more widely available than direct microstructure observations. I then applied the optimised models to two ship-based CTD sections in the Brazil Basin to investigate spatial and temporal variability in turbulent mixing.
“I also gained extensive hands-on experience in scientific programming and data analysis, working with tools and packages including NumPy, pandas, xarray, SciPy, scikit-learn, XGBoost, PyTorch and Matplotlib. More broadly, the internship gave me valuable experience of what day-to-day research is actually like. In particular, I learnt how to interpret results that do not always behave as expected, think critically about possible explanations, and identify sensible next steps. Overall, the experience has strengthened both my technical skills and my confidence and interest in pursuing research further."
Hannah Scothern
Project: Constraining a terrestrial ecosystem model using greenhouse gas flux and remote sensing data
“My internship focused on constraining an ecosystem model, which provided an excellent opportunity to apply the skills and practices I acquired during my Natural Sciences IB Quantitative Environmental Science coursework.
“Because terrestrial ecosystems were a completely new subject area for me, the project provided a steep but rewarding learning curve in ecosystem dynamics. It was also great to be based within the Centre for Landscape Regeneration and see how interdisciplinary environmental research operates in practice.
“Attending the ICCS Summer School was also immensely valuable, as it provided structured training in research software engineering and computational best practices which proved extremely useful for executing the technical aspects of the work.
“My supervisor was incredibly supportive throughout, providing the guidance needed to keep the research moving efficiently. I was amazed by how much progress we made in the eight weeks, culminating in a manuscript we are currently preparing for publication.”
Ameer Alhashemi
Project: Geospatial Foundation Models
“During my internship, I looked at whether TESSERA, a model trained on satellite imagery, could be useful for understanding the structure of forests. I compared satellite data with detailed airborne laser measurements from forests in Scotland and the Netherlands. I found that TESSERA was quite good at picking up things like how rough or open the canopy was and differences in tree cover, although estimating exact tree height was more difficult. I also tried out a few different machine-learning approaches and learned a lot about how to fairly test models across different locations.
“One of the things I enjoyed most was having the freedom to explore results that seemed interesting and let the project develop from there. I regularly talked through ideas with David, Aneesh and the wider team, which made the project feel very collaborative. I also got a lot of practical experience with environmental machine learning, remote sensing and working with large geographical datasets.
“Future interns can expect to work on real research problems, try out their own ideas and see how a project changes as new results come in. I had a lot of fun during the internship and learned a huge amount, and it definitely made me want to keep exploring work that combines computing and environmental science.”
Tarun Nandi
Project: Accelerating UKCA using data-driven timestep adaptation
“During my internship at ICCS, I used machine learning to improve the computational efficiency of UKCA, an atmospheric chemistry model used in chemistry research. My project explored whether neural networks could predict when the model needs smaller timesteps to solve its chemistry equations, aiming to reduce computational cost while sustaining accuracy.
“I obtained valuable experience in preparing data, designing experiments, and evaluating models, along with a better understanding of the challenges of applying machine learning to more scientific problems. Attending the ICCS summer school was another highlight where I learnt so much and was introduced to many ideas beyond my own project. Overall, I really enjoyed working with the research engineers across ICCS, and the experience definitely strengthened my interest in pursuing a career in research engineering.”
John Zhen
Project: Symbolic execution of Fortran programmes
“Over the summer, I designed and developed a symbolic executor for Fortran, which uses automated reasoning to explore execution paths and verify program properties. As Fortran is a language that is still widely used in scientific computing, this project can help identify bugs and improve confidence in software that real scientific results depend on.
“One of the things I enjoyed most about the internship was having the freedom to take ownership of the project while still being able to discuss ideas and problems with my supervisors, Nikolaus and Dominic. It was incredibly rewarding to see the project grow from an initial prototype into something capable of handling increasingly complex language features.
“The experience has strengthened my interest in programming languages and scientific computing research, and has encouraged me to further explore these areas in the future. Furthermore, everyone at ICCS was incredibly welcoming and supportive, which made the internship a genuinely enjoyable experience, and I would strongly encourage future students to apply for the programme.”
Thank you!
We’d like to congratulate all our interns and also thank their supervisors and co-supervisers for their help and support:
- Luke Abraham, Department of Chemistry/NCAS
- Laura Cimoli, ICCS
- David Coomes, Centre for Landscape Regeneration
- Andrew Friend, Centre for Landscape Regeneration
- Nikolaus Huber, ICCS
- Robert Rouse, ICCS
- Joe Wallwork, ICCS