Research Software Engineer

About us:

CogStack ( https://cogstack.org/) is an award winning ecosystem of tools and workflows that facilitate the ingestion, structuring, organising and visualisation of Electronic Health Record (EHR) data built by a multidisciplinary team of software developers, machine learning engineers, clinical researchers and health informaticians.

The CogStack team benefits from sitting within a leading programme of clinical, health and bioinformatics at the South London and Maudsley (SLaM) Biomedical Research Centre (BRC) and forms a key component of both the Centre for Translational Informatics ( www.ctiuk.org) and actionable analytics theme of the recently awarded Health Data Research UK (HDR UK) London site.

Major funding has been awarded by the Office for Life Sciences, InnovateUK and recently a Stage 3 AI for Health and Social Care Award from NHSx. The ecosystem has already been recognised in Government reports to the Chief Medical Officer, NHSx AI report, NHS Tech Plan and keynote speeches by the Health Secretary.

About the role:

The postholder will contribute to the technical development, deployment and maintenance of the ecosystem, which includes the CogStack-Nifi ( https://github.com/CogStack/CogStack-Nifi) and MedCAT (https://arxiv.org/abs/1912.10166 ;  https://github.com/CogStack/MedCAT) projects alongside any deployment specific enhancements and specialisations. This work will include contributing to our clinical NLP tools, algorithms and interfaces used by clinical specialists.

The post holder will be expected to contribute to:

• Extend our software platforms for electronic healthcare record research and direct care that includes Natural Language Processing / Large Language Model (LLM) capability, e.g. extracting structured information from free text of both biomedical literature and electronic patient records exploiting HPC, including GPUs embedded within NHS infrastructure.
• Development and deployment of ML operations software and tooling for ML / LLM algorithms working over free-text clinical data and potentially mixed modalities (waveform, DICOM images etc.)
• Data manipulation and analysis capability, e.g. machine/deep learning in R/Python or other available tools with an interest in integrating multimodal data from EHRs
• Excellent communication skills are required to describe and understand the requirements, key challenges and evaluate methodologies of electronic health records research and development.

The successful candidate will have a MSc or equivalent experience in a relevant discipline (e.g., computer science, clinical informatics, software development, knowledge representation, machine learning, text mining). Relevant data analysis and programming skills (e.g. Python, R) are required.

This is a full-time post (35 hours per week), and you will be offered a fixed term contract until 30th July 2026.

About you:

Essential criteria

• MSc or equivalent experience in relevant area
• Programming (one or more of the following: Python, JavaScript, Java, C#)
• Knowledge and experience of Software development best practices
• Experience working with multi modal clinical data such as clinical record text, structured data, and/or clinical imaging
• Use of Container technologies such as Docker, Podman, K8s in production environments
• Experience working with deep learning models across multiple modalities

Desirable criteria

• MLOps experience of with multi modal deep learning or classical machine learning models in a production setting
• Experience of biomedical semantic ontologies
• Knowledge and experience of web application development with frameworks such as React, Vue, Next

Advert information