Open positions

Background

Hospitals routinely acquire multiphase CT scans to diagnose liver cancer, capturing rich spatial detail and time-resolved contrast enhancement patterns. Yet much of the quantitative information in these scans—such as texture, shape, and phase-to-phase enhancement behaviour—is not fully exploited in routine clinical reporting.

Developing AI methods that genuinely generalise across hospitals and scanning protocols depends on strong foundations: clean, well-curated datasets and a clear understanding of what the data contains (and what it doesn’t). In this internship, you’ll work with a recently acquired dataset from a joint project led by Nottingham University Hospitals NHS Trust, with international collaborators, combining multiphase CT imaging with detailed clinical data.

Host and placement

• Hosted in the team of Dr George Gordon (OPTIMlab), University of Nottingham
• Close clinical collaboration via Nottingham University Hospital’s NIHR Nottingham BRC Liver & GI disorders theme
• Up to 10 weeks during summer 2026

What you’ll do (core objectives)

You’ll help prepare, validate, and prototype analysis pipelines for a real-world multiphase liver CT cohort:

1. Explore and understand the dataset
   • Map cohort structure, scanning protocols, and imaging and clinical variables.
2. Build an EDA + validation toolkit
   • Implement systematic checks such as patient–scan linkage integrity, missing phases, scan parameter variability, and corrupted series detection.
   • Produce a clear data readiness report.
3. Create AI-ready representations
   • Evaluate practical options for generating embeddings/latent representations (including pre-trained and 3D representation-learning approaches where useful).
   • Support later tasks such as classification or prognosis.
4. Stretch goal (optional, time permitting)
   • With clinical collaborators, prototype end-to-end AI models that can extract clinically relevant outputs from multiphase scans.

What you’ll learn

• How real clinical imaging data is structured (DICOM basics, multiphase CT conventions, common pitfalls)
• Building robust, reproducible data pipelines (validation, documentation, and QA for medical AI)
• Practical dataset auditing and EDA for medical imaging research
• Modern AI tooling for 3D representation learning and pre-trained modelling
• How to work effectively in an interdisciplinary team (AI/engineering/physics + clinicians)

Who we’re looking for

We welcome applications from undergraduates who are excited by applied AI in healthcare. You might be a good fit if you have:

Essential

• Strong coding skills (data handling, plotting, basic software engineering practices)
• Interest in medical imaging and/or machine learning

Helpful (not required)

• Familiarity with medical imaging data and concepts
• Experience with ML libraries (PyTorch, scikit-learn)

Support and environment

You’ll join OPTIMlab, an interdisciplinary research team developing next-generation ultra-thin imaging devices and AI-enabled reconstruction (more details at www.georgesdgordon.com). You’ll be supported by Dr Gordon, clinical collaborators, and a team with active AI expertise and access to strong compute resources.

Pay and practicalities

• 36.25 hours/week for 10 weeks at c. £12.78/hour
• Includes a £500 research consumables budget to support the project
• Internships must take place between Monday 22 June and Friday 18 September 2026
• Selection and appointment are subject to standard eligibility checks under the EPSRC-funded vacation internship scheme

Widening participation

We are especially interested in hearing from candidates from underrepresented backgrounds. If you’re unsure whether you meet the scheme’s eligibility criteria, please get in touch to discuss the full details.

How to apply

To express interest, please email george.gordon@nottingham.ac.uk with:

• A short CV (1–2 pages)
• A brief statement (max 1 page) describing relevant skills/experience and why this project interests you