The Undergraduate Research Opportunities Programme (UROP) enables Cambridge undergraduates to spend 8-10 weeks over the summer helping with research activities taking place at Cambridge University.
The paid internships offered through this programme give students insights into the research being carried out by the world-class academics here. They also give students opportunities to develop some of the technical and transferable skills required in research activities and many other graduate careers.
A list of the UROP placements being offered in this Department will be advertised here as soon as they become available. You can choose to apply for any of those on offer, including those offered outside this Department (i.e. the placements offered in the Engineering Department).
If you have any questions about how the UROP programme operates here in the Department of Computer Science and Technology, please email urop-admin@cst.cam.ac.uk.
(If you are a member of staff interested in running a UROP project, please see our internal information pages.)
How do UROP placements work?
The programme is open to any Cambridge University undergraduate who is not in their final year of study. A UROP placement takes place over the Long Summer Vacation and normally lasts 8-10 weeks. UROP students will be paid National Living Wage, plus the appropriate holiday pay. Students' pay is subject to appropriate deductions (i.e. for income tax and National Insurance contributions).
Before projects start
If your application is successful and you are assigned to a project, you will be sent a number of forms to complete and sign. These must be returned to our Human Resources (HR) team before the project starts.
Your schedule of work should also be negotiated with your supervisors before starting the project. Your supervisor is also responsible for setting you up on the Department's visitor database so that you will be able to access the research side of the building.
Requirements
Please be aware that students on UROP placements MUST be living in the UK (even if they are working from home) and must have the right to work in the UK. They are expected to work on the project 37.5 hours per week.
Weekly timesheets will be submitted through Cambridge Casual Worker System (CCWS). Ask your supervisor to confirm your timesheet to our Human Resources (HR) team each week.
Students' work is expected to conform to an acceptable technical standard. Supervisors retain the right to terminate appointments, if performance remains below the required standard, following due warning. Students should also note that their UROP will be terminated immediately if they cease to be registered for the following academic year (e.g. by failing in the Tripos examinations).
Available UROP Projects - updated 18/05/2026
Project 1
Title: Learning the Language of the Human Body: Multi-Modal AI from Wearable Data
Supervisor: Dong Ma - dm878@cam.ac.uk
Essential knowledge, skills and attributes: Python, biosignal processing, unsupervised learning
Description of project:
Wearable devices continuously collect human signals and recent advances in AI have led to powerful "foundation models" for biosignals (like ECG or EEG). However, most existing models: (1) Focus on a single type of signal (e.g., only heart activity); (2) Learn patterns without understanding what they mean physiologically; (3) Combine multiple signals in simplistic ways, missing how body systems truly interact. This project aims to move beyond that--toward models that capture how different physiological systems work together by developing multi-modal foundation model for physiological signals. You will work with the supervisor and contribute to developing a multi-modal, physiology-aware representation learning pipeline with a breakdown approach designed by the supervisor. By the end of intern, you will gain hands-on experience with modern AI techniques, insight into interdisciplinary research combining, and a strong project you can showcase in future applications or research opportunities.
Length of project: 8 weeks
Remote or in person (student must be in the country either way): both are acceptable
Project 2
Title: Scaling Neuro-Symbolic Grammar Induction
Supervisor: Dr Weiwei Sun - ws390@cam.ac.uk
Co-supervisor: Yuan Gao (fourth year PhD student)
Essential Knowledge, Skills and Attributes:
* Good programming skills in C/C++ and Python
* Basic understanding of neural networks and deep learning
* Familiarity with machine learning and probabilistic modelling is desirable
* Interest in natural language processing
Description of Project:
Grammar induction is a fundamental problem in natural language processing and computational linguistics. The goal is to automatically learn grammatical structure from linguistic data. Existing approaches can broadly be divided into two paradigms: traditional statistical methods based on symbolic grammar formalisms, and neural approaches based on deep representation learning. Combining symbolic grammatical constraints with neural models can substantially improve interpretability and structural generalisation. However, such neuro-symbolic approaches introduce a highly challenging deep latent variable learning problem. In particular, the symbolic component typically requires dynamic-programming-based inference over large structured spaces, of which parallesing it on GPU is extremely difficult. This project aims to develop a variational inference framework for scalable neuro-symbolic grammar induction. The core idea is to separate the symbolic inference procedures from neural parameter updates as much as possible. Specifically, the project will investigate methods that decouple dynamic-programming-based structured inference from backpropagation-based optimisation, thereby improving computational efficiency and enabling substantially larger models and datasets to be used.
Length of Project: 8 weeks
Mode of Supervision: Hybrid. Dr Sun may be away from Cambridge in August, and some supervision meetings may therefore take place remotely.
Project 3
Title: MAKOTO in English
Supervisor: Katie Seaborn - kas214@cam.ac.uk
Description of Project:
I created the first "older adult"-voiced virtual assistant in Japan, in Japanese, from scratch. That's because most base models and voice clips are English ... so we had to build a Japanese "older adult" text-to-speech (TTS) system from scratch. The purpose was to use it as a research instrument: a positive "older adult" stimulus for an implicit bias intervention (longitudinal personalized storytelling with ChatGPT under the hood). Now I want to run the study in English. Also, it's been several years and there seems to be free English "older adult" TTS voices out there. I'd like a student to (a) hook up an English voice and (b) translate the entire system to English (user interface, prompts, some cherry-picked voice samples). If there's time, I'd like the student to set up an English version of the web-based/online IAT, which measures implicit bias in the study. There's some free code via GitHub by the Harvard team and others; we also have a Japanese version that could be reverted.
Skills: Knowledge of web APIs; knowledge of TTS or related technologies; knowledge of LLM APIs like the OpenAI API; some knowledge of audio editing (e.g., cherrypicking, Audition or similar software); knowledge of web programming (e.g., JavaScript, HTML, CSS, PHP); knowledge of data formats (e.g., JSON); basic web server knowledge (e.g., Node.js)
Length of Project: 8 weeks
Project 4
Title: Bilt - Novel low order physical systems modelling tool
Supervisor: James Emberton
Department: Aviation Impact Accelerator, Whittle Lab, Dept of Engineering
Email Address: je484@cam.ac.uk
Project Description
Bilt is a novel low order computational modelling tool. We are developing a completely novel approach to support analysis of complex physical systems, in an accessible way, based on functional programming principles. This project is based at the Aviation Impact Accelerator at the Whittle Lab and supports work to de-carbonise aviation. However, we anticipate that the Bilt tool can be applied more generally in the world of systems modelling.
A number of potential projects are available depending on the student's skills and interests. These could include:
- design and evaluation of novel computational and static analysis infrastructure
- prototyping new physical systems toolboxes
- developing cloud compute and data workflows
This is an opportunity to help shape the future of a novel open-source computational tool with enormous potential.
Essential Knowledge, Skills, and Attributes
All projects require strong familiarity with Python as a minimum and a strong interest in sustainable aviation.
Depending on the project the following skills may be useful:
- Programming languages such as Rust, OCaml, or Haskell
- Knowledge of computational graphs, compilers, computer algebra, or functional programming
- Array computation (e.g. JAX), compiler infrastructure (e.g. LLVM), or symbolic mathematics (e.g. CasADi)
- A background in Chemical, Mechanical, Electrical, Aerothermal Engineering or Computer Science
- Interest in open-source software development
Timing
Applications will be assessed on a rolling basis.
Any applicants will be expected to start their 8–12-week placement in July
Continuation Opportunities
There is potential to support continuation work for undergraduate thesis or MSc projects.
Supporting Information: https://aiazero.org/
Application Details
Name and email of the person receiving the applications: Anna Petrosyan ap2522@cam.ac.uk
Deadline for applications: Rolling
