From Molecules to Meaning: Agentic AI for Health
Dr. Robail Yasrab is a Senior Researcher at the MRC Biostatistics Unit, University of Cambridge and a Visiting Researcher at
Oxford’s Institute of Biomedical Engineering. He develops agentic AI and biomedical foundation models
that fuse multi-omics, genomics, clinical text, and imaging into explainable decision support and deployable diagnostics.
Agentic AI
Biomedical FMs
Multi-omics
Clinical NLP
Medical Imaging
Explainability
Grants & Programmes
KAMAL Health — Knowledge-Augmented Multi-Modal Arabic LLMs for Healthcare
Lead/Coordinator. 1.5M AED Dubai Future Foundation award (2025–). Building bilingual (Arabic/English), knowledge-augmented medical LLMs.
Dubai RDI ·
University of Birmingham ·
Ruchit Agrawal
Wellcome Leap — In Utero (2023–)
Contributing to scalable measurement and modelling in pregnancy.
wellcomeleap.org
Key Projects
PULSE
ERC-funded programme modelling expert sonographer behaviour using ultrasound, eye-tracking, and probe-motion data to power next-gen assistive interpretation.
Project page
RootNav 2.0
Co-created deep-learning tool for plant phenotyping; open-source and widely used.
GitHub
Resume
Experience
Senior Researcher · MRC Biostatistics Unit, University of Cambridge
2023 — Present
Visiting Researcher · Institute of Biomedical Engineering, University of Oxford
2023 — Present
Postdoctoral Research Assistant · Biomedical Image Analysis Group, University of Oxford
2020 — 2023
Research Fellow · Computer Vision Lab, University of Nottingham
2018 — 2020
Postdoctoral Researcher · USTC
2017 — 2018
Education & Training
University of Oxford, UK
2021 — 2022 · Researcher Strategy Consultancy Programme
USTC, China
2014 — 2018 · PhD in Advanced Machine Learning · CAS-TWAS President’s Fellowship (2014)
Gothenburg, Sweden
2012 — 2014 · MSc in Computer Science & Engineering (Algorithms & ML)
Selected Publications & Open-source
Automated Analysis of First-Trimester Ultrasound
IEEE Transactions on Medical Imaging. End-to-end modelling of fetal ultrasound video, workflow, and eye-tracking.
Paper · PMC
RootNav 2.0 — Deep Learning for Plant Phenotyping
GigaScience. Widely used open-source tool for automatic navigation of complex root architectures.
Paper · GitHub
Also published in Medical Image Analysis, ECCV/ISBI. Teaching: Medical Image Processing, Signals & Systems/DSP, Embedded ML, AI for Biomedical Systems.