Vacancies

Chaste is an open-source research software project originally developed at the University of Oxford for simulating biological systems, particularly collections of cells and tissues. It is widely used in computational biology research, with applications ranging from cancer modelling to heart simulations and tissue development. The project is written mainly in C++ and places a strong emphasis on high-quality, sustainable scientific software. 

One of Chaste’s strengths is its flexible framework for simulations of cell populations. Researchers can represent tissues in several different ways and study how cells grow, move, divide, and interact over time. Because of this, Chaste is well suited to combining experimental data with computational modelling. 

In recent years, there has been growing interest in using microscope images to generate realistic starting conditions for simulations. Advances in AI image analysis now make it possible to automatically identify and segment cells within biological images. Integrating these techniques with Chaste creates opportunities to build simulations directly from experimental observations, helping bridge the gap between biological experiments and computational models. 

Project Description

This project will extend ongoing work linking AI-based image analysis with Chaste simulations. Existing tools can already segment some microscope images of cell populations and use these to create initial configurations for some simulations. The aim of the internship is to make this workflow more useful, flexible, and scientifically informative. 

The project can be shaped around the student’s interests. One direction would focus on improving the image-analysis pipeline itself: testing modern AI segmentation tools, improving robustness on difficult biological images, and developing clearer workflows for converting image data into simulations.  

A second direction would focus more on scientific software development in C++. This could include building tools that convert segmented image data into simulation geometries compatible with different Chaste cell-population models and improving support for importing experimental data into simulations. 

There is also scope for exploratory scientific work, such as investigating whether simulations initialised from real biological data behave differently from simulations started from idealised synthetic arrangements. 

Searching for periodic radio signals from neutron stars (aka pulsars) in the Galaxy is a key scientific pursuit in radio astrophysics. In pulsar surveys, large volumes of time-series data are processed to identify possible periodic signals originating from pulsating neutron stars. These candidates must then be “folded”, (combined coherently using their trial period and related parameters), to enhance the signal and assess whether they are likely to be truly astrophysical sources. Folding is therefore a critical step in candidate confirmation, but it can also be computationally demanding, particularly in large-scale searches where many candidates must be processed rapidly. This calls for the investigation of faster folding algorithms that can be deployed on accelerators like Graphics Processing Units (GPUs). 

Project Description

The main aim of the project is to investigate whether GPUs can be used to accelerate candidate folding and make this stage of the search pipeline more efficient. The student will learn the scientific background to pulsar searches and the practical role of folding in pulsar candidate analysis, while also gaining experience in GPU programming. In particular, they will be introduced to CUDA and the development of GPU kernels for scientific computing applications. The project is flexible enough to accommodate different programming backgrounds and interests, and the student may choose to develop and deploy these GPU algorithms in either C++ or Python. A key component of the project will be benchmarking the GPU-based implementations against existing CPU approaches in order to quantify potential speed-ups and identify any trade-offs in accuracy, flexibility, or resource use. 

By the end of the internship, the student will have gained experience in astrophysical data analysis, scientific programming, GPU computing, and performance evaluation. The overall goal is to assess whether GPU-based folding algorithms could be deployed effectively in real pulsar search systems, particularly those operating in high-data-rate or real-time environments.

Gutenberg is OxRSE’s open-source teaching platform for research software engineering and scientific computing, available both as a self-hosted web application and through Oxford’s deployment at https://train.rse.ox.ac.uk. The platform hosts a large collection of teaching material written in Markdown, covering topics such as programming, version control, testing, reproducibility, and data analysis. 

Teaching material can be delivered either through self-paced “Courses” or through institutionally organised “Events”, where online lessons are combined with lectures, workshops, and tutor support. Gutenberg therefore acts both as a teaching platform and as an interactive learning environment. 

A key feature of Gutenberg is its built-in feedback system. During Events, students can leave comments directly within lessons and provide feedback on individual Problems and Solutions embedded in the material. This has proved valuable for identifying confusing explanations, stale content, and broken exercises, helping organisers continuously improve courses over time. 

However, the current feedback tooling remains limited. Feedback is mainly consumed during the Event itself, with little support for retrospective analysis, summarisation, or long-term reporting. In addition, post-Event surveys are generally collected through external tools, making it difficult to combine feedback sources or analyse trends across multiple deliveries of the same course. 

Project Description 

This project will improve Gutenberg’s feedback and reporting system, making it easier for organisers to understand how students interact with teaching material and identify areas for improvement. 

One part of the project will focus on integrating post-Event feedback directly into Gutenberg. The student will design and implement a dedicated survey system within the platform, allowing organisers to collect structured feedback after an Event without relying on external tools. 

Another part of the project will focus on improving visibility and analysis of existing feedback. Currently, Problem-specific feedback is stored but difficult to review at scale. The project will develop tools to summarise this information, including statistics such as average ratings, counts of responses, and collections of written comments. These summaries should be accessible through the web interface and exportable to common file formats for later analysis. 

The project can be shaped around the student’s interests. A student more interested in web development and user experience could focus on designing clear interfaces and workflows for viewing and managing feedback. A student more interested in backend or software engineering work could focus on data models, APIs, exporting functionality, and analytics pipelines. 

There is also scope for exploratory work around educational data analysis, such as identifying patterns in student engagement, understanding which exercises consistently cause difficulty, or evaluating how feedback evolves over repeated deliveries of the same teaching material.