Martin
Chapman

Working on the secure deployment of Large Language Model (LLM) technology that converts unstructured NHS clinical documents into analysable data at scale. This represents one of the largest deployments of this technology in the health service. Leading external trust deployment, and developing training, deployment and inference pipelines to support the work. Example models include GenoLlama (genomic biomarkers) and OncoLlama (cancer).

Led on the development and deployment of a RAG-based chatbot that provides an interactive frontend to NHS online genomic notes for clinicians (GeNotes).

Led on the rapid development of a model that combines AI simulation techniques, big data and PPIE to explore the impact of cold home interventions during the cost-of-living crisis. Work covered by Sky News, The Independent and Bloomberg, and findings fed back into government policy.

Co-led the development of a dashboard to enable stroke patients to access their data, to support research into data-based interventions. Patients recruited via the South London Stroke Register (SLSR), a dataset jointly sponsored by King's College London and Guy's and St Thomas' NHS Foundation trust.

Led on the exploration of how sensor data can be used to better support patients with cardiometabolic disorders. Worked with Metadvice Ltd., a global health technology company, to deliver additional wearables-based components for their cardiometabolic decision-support system (DSS), which has been rolled out across the UK.

Software01

reflectMiddleware for interacting with wearable devices

Stack

Worked as a part of HDR UK's National Phenomics Resource team to construct the UK's first national repository of computable disease definitions (phenotypes). Led on the development of Phenoflow, software that standardises repository phenotypes as Common Workflow Language (CWL) workflows.

Publications04

Generating Computable Phenotype Intersection Metadata Using The Phenoflow Library2025 · AMIA Joint Summits on Translational Science
M Chapman, L Rasmussen, J Pacheco, L Wiley and V CurcinConnecting Computable Phenotypes With Multiple Health It Standards Using The Phenoflow Library2022 · AMIA Clinical Informatics Conference
M Chapman, L V Rasmussen, J A Pacheco and V CurcinPhenoflow: A Microservice Architecture For Portable Workflow-Based Phenotype Definitions2021 · AMIA Joint Summits on Translational Science
M Chapman, L Rasmussen, J Pacheco and V CurcinUsing Computable Phenotypes In Point-Of-Care Clinical Trial Recruitment2021 · Digital Personalized Health and Medicine
M Chapman, J Domínguez, E Fairweather, B C Delaney and V Curcin

Software01

phenoflowStandardise and share computable disease definitions

Talks06

Phenoflow: An Architecture for FAIRer PhenotypesThe Northwestern University Clinical and Translational Sciences (NUCATS) Institute/Institute for Augmented Intelligence in Medicine (I.AIM)/Center for Health Information Partnerships (CHIP), Northwestern University, Chicago, 2025 Generating Computable Phenotype Intersection Metadata Using the Phenoflow LibraryAMIA Informatics Summit, Pittsburgh, 2025 (presented by Professor Vasa Curcin)Phenoflow: An Architecture for Computable PhenotypesApplying FAIR principles to computable phenotype libraries, ELIXIR All Hands, Amsterdam, 2022Phenoflow 2021Clinical Natural Language Processing Group, University of Edinburgh, 2021Phenoflow: A Microservice Architecture for Portable Workflow-based Phenotype DefinitionsAMIA Informatics Summit, 2021Using computable phenotypes in point of care clinical trial recruitmentMedical Informatics Europe (MIE), 2021

Stack

Worked as a part of an interdisciplinary team to research the feasibility of employing a decision-support system (CONSULT) to enable chronic disease self-management. Led the development of a set of microservices for the DSS to support wearables data collection and integration, AI reasoning and real-time data analysis.

Publications01

Computational Argumentation-Based Clinical Decision Support2019 · 18th International Conference on Autonomous Agents and MultiAgent Systems
M Chapman, P Balatsoukas, M Ashworth, V Curcin, N Kökciyan, K Essers, I Sassoon, S Modgil, S Parsons and E I Sklar

Software01

consultModular decision-support system for chronic condition self-management

Stack

Involved in the development of a distributed ledger PoC for a major Fortune 500 financial institution. Focused on the pre-processing of raw financial data for storage on an Eris (Monax) chain.

Involved in exploring how finite automata can be used to represent software error traces. Developed a mechanism to estimate model-checking bounds, and the project’s merge feature. Sponsored by Google.

Software01

learning-errorsAbstractly represent software error traces as finite automata.

Stack