Project Proposals

Below is a list of internship projects which are continually being updated and added to. We also welcome project proposals from both prospective students and interested organisations. Please contact us at our mailbox if you would have queries or would like to discuss a proposal.

Open Projects Accepting Applications

These projects are the list that our 2026 cohort could apply to. Projects for 2027 will be confirmed and appear here in early August. We are happy to discuss the briefs and hear suggestions to fine tune them.

• The NHS Patient Profile - Representations for Clinical AI

  MultiModal Learning Representations MultiModal

• Evalsense continued

  Explainability Benchmarking Text

• Fact-Level Privacy Leakage Detection and Mitigation

  PETs Memorisation Text

• Mechanistic Interpretability for AI Systems in Healthcare

  Explainability Circuits MultiModal

• NHSSynth Studio

  Synthetic Framework Tabular

• TailSynth - Rare Event Synthetic Generation

  Synthetic RareEvents Tabular

• Detecting LLM Model Collapse

  Explainability LLMs Text

• Sparse Networks for Explainability

  Explainability SparseNetworks MultiModal

• Linguistic Robustness in Triage

  NLP Robustness Text

• Foundation Models for Clinical Image Analysis

  Machine Learning Foundation Models Images

• Uncertainty Quantification for Clinical AI

  Explainability Uncertainty Tabular

• Fairness and Equity in Clinical AI

  Machine Learning Fairness Tabular

• Longitudinal Patient Trajectory Modelling

  Machine Learning Temporal Tabular

Closed Projects

Paused Projects

Paused projects are ones which we have advertised in the past and still interested in but either have been de-prioritised, need further research or there is an issue with data access delaying the work. Happy to discuss these briefs if you see opportunity.

• Domain Gap Bridging

  MachineLearning VAE Images

• Adversarial Attacks to Quantify Privacy of Synthetic Data

  Synthetic Adversarial Tabular

• Reinforcement Learning as a mechanism to improve alignment and control in Multi-modal models

  NLP ReinforcementLearning MultiModal

• Structural Topic Models and Sentiment Model Robustness - Extension

  NLP STM Text

• Graph-based representations and techniques for Healthcare Data

  Graphs Longitudinal Tabular

• Building towards a Primary Care Digital Twin

  Synthetic DigitalTwin Tabular

• Augmenting Text Generation in Healthcare with Knowledge Retrieval Approaches

  NLP GenerativeLanguage Text

• Automated Text Descriptions from Imaging - Next Steps

  NLP Embeddings MultiModal

• Causal Models to Generate Synthetic Longitudinal Data

  Synthetic DAGs Tabular

• Exploring Large-scale Language Models with NHS Incident Data - Next Steps

  NLP Embeddings Text

• Fairness in Healthcare Data and Models

  Explainability Equity Tabular

• Synthetic Data Exploration - Probabilistic Graphical Models

  Synthetic BayesianNetworks Tabular

• Exploiting an NHS Monitor Corpus

  NLP MonitorCorpus MultiModal

• Extending NHSSynth into Multi-table, Multi-modal, and Longitudinal Data

  Synthetic VAE Tabular

• Exploring Human-Computer Interaction and Collaboration in Healthcare Settings

  Explainability DecisionSupport Tabular

• Exploring Digital Watermarking and Usage-Aware Agents for Healthcare AI Governance

  PETs Agents Tabular

• Evaluating the risks and value of multiple geospatial data linkages

  Linkage Evaluation Tabular

• Pharmacy Prioritisation Support

  MachineLearning Pharmacy Tabular

Completed Projects

Completed Projects are ones which our previous interns have finished and reported on. There may still be many areas of investigation possible with these and so we are happy to talk about redeveloping or continuing.

• Applied Privacy Enhancing Technologies in Healthcare

  PETs DifferentialPrivacy Tabular

• Automatic Speech Recognition with Summarisation Evaluation

  Explainability ASR Audio

• Building towards a Patient Record Digital Twin

  Synthetic DigitalTwin Tabular

• Fairness in MultiModal Healthcare Models

  MachineLearning AlgorithmicFairness MultiModal

• Genomic Data Incorporated into Healthcare Predictions

  MachineLearning PrecisionMedicine GenomicData

• Simulating Healthcare Scenarios with LLMs as Agents

  Simulation ABM Text

• Causal Inference to Estimate the Impact of Interventions

  MachineLearning CausalInference Tabular

• Automated Evaluation and Benchmarking of Data Linkage Quality

  Linkage Evaluation Tabular

• Generative AI Evaluation

  NLP Evaluation Text

• Understanding Bias in Multi-modal models within Healthcare - Continued

  Explainability Fairness MultiModal

• Advancing Privacy Fingerprint (PrivFp) using modern generative methods

  Synthetic Privacy Text

• Predicting Demand to Support Health Inequalities – working with the Leeds Data Model

  MachineLearning Classification Tabular

• System testing of AI technologies – working with the HRA

  Explainability AISafety MultiModal

• Process Mining to Generate Healthcare Pathways

  Simulation ProcessMining Tabular

• Explaining facial skin disease classification using LIME

  Explainability LIME Images

• Predicting the Impact of Health Inequalities - Diabetes

  MachineLearning Classification Tabular

• Structural Topic Modelling on Contact Tracing Feedback Data

  NLP STM Text

• Synthetic Data Exploration - Variational AutoEncoder for Healthcare Data

  Synthetic VAE Tabular

• Value of Commercial Product Sales Data in Healthcare Prediction

  MachineLearning ModelClassReliance Tabular

• Synthetic Text Generation in Healthcare

  NLP SyntheticGeneration Text

Archived Projects

Projects which we have decided not to go ahead with either due to lack of data or deprioritised work.

• Building an NHS Agent Based Model in HashAi

  Simulation ABM Tabular

• Evaluating the Implementation of Automated Clinical Coding on Neurology Clinic Letters

  NLP NER Text

• Granular Mapping

  Geospatial Curation Tabular

• Inequalities in Population Health Data - Cardiovascular and Heart Disease

  MachineLearning Classification Tabular

• Metadata API for ABMs

  Simulation ABM MultiModal

• NHS GreenSpace

  Sustainability ImageSegmentation Images

• NHS Semantic Search

  NLP SemanticSearch MultiModal

• Prediction of winter respiratory admissions – working with SNEE ICS

  MachineLearning Prediction Tabular

More information on the internship scheme can be found on the NHS England internship scheme for innovation and analytics in health website.