PhD researcher · University of Southampton
I’m Inam Ullah Khan, a final-year Computer Science PhD researcher. I develop machine learning and statistical methods that connect retinal imaging with molecular and clinical data to study systemic health.
Exploring academic and industry research opportunities
Research, publications & conferences
Featured in Forbes
· William A. Haseltine
The Forbes article “AI Eye Exams Could Detect Heart Disease, Diabetes And More Earlier” links to our survey, The Eye as a Window to Systemic Health, in its discussion of retinal imaging and chronic disease research.
Connecting retinal imaging, artificial intelligence, and the wider conversation about systemic health.
Our study of cross-sectional retinal–lipidomic associations in a generally healthy biobank population has been published in Scientific Reports, part of Nature Portfolio.
Read the paper
Our work on candidate biomarkers for cardiovascular health is available as a Springer chapter in the MICAD 2025 proceedings.
Read the chapter
The Dual-Edge Spatial-Jacobian image graph paper was accepted in the Emerging Research track.
Conference details
Our framework studies causal modelling with fragmented biomedical data.
View the paperI began my professional journey in civil engineering as an associate engineer, learning to turn complex problems into practical solutions. Training in physics and mathematics gave me a language for understanding systems; computer science gave me the tools to model them.
Today, these perspectives come together in my PhD research. I develop machine learning and statistical methods that integrate retinal imaging with molecular and clinical data, seeking interpretable insights into human health. Engineering keeps my questions grounded in practical constraints; mathematical reasoning helps me examine the assumptions behind the answers.
Engineering practice Mathematical reasoning Computational health research
Selected research
Biomedical measurements are often incomplete across datasets. RetiSEM studies how to recover interpretable causal structure from fragmented, multimodal evidence.
Causal models · Multimodal integration
Explore RetiSEMCan the eye reveal molecular signals elsewhere in the body? This study characterises cross-sectional associations between retinal vascular traits and lipidomic profiles in a generally healthy biobank population.
Retinal phenotyping · Lipidomics · Statistics
Read the studyThe Dual-Edge Spatial-Jacobian framework represents vessel–lesion spatial relationships and lesion–biomarker relationships as an image graph for diabetic retinopathy grading.
Image graphs · Computer vision · Interpretability
Read the frameworkSelected publications
A common thread runs through my work: use mathematical structure and computational methods to make biomedical data more interpretable, from robust image models to molecular associations and causal reasoning.
Mathematical structure: causal modelling for the practical challenge of incomplete, multimodal biomedical evidence.
Connecting biological scales: statistical analysis of cross-sectional associations between retinal vascular traits and molecular profiles. Published on 27 October 2026 in Scientific Reports.
A systems perspective: a survey connecting retinal imaging methods with the study of systemic health.
Practical modelling foundations: studying ensemble diversity for medical image classification under data scarcity and imbalance.
Methods & experience
My work brings together medical image analysis, statistical modelling, and biomedical data integration. I also support programming, data analytics, and image processing teaching at Southampton.
Experience and technical backgroundWhat drives the work
Growing up in Bajaur, Pakistan made access to healthcare a personal concern. It continues to shape the questions I ask about how computational research can contribute to understanding human health.
More about my backgroundWriting & resources
I write about mathematical ideas, computational research, and practical workflows, including a guide to running retinal image analysis with AutoMorph locally and on HPC systems.
Explore the research notebookThrough my Educational Hub on Facebook, I share fully funded scholarship opportunities and practical guidance on preparing application emails, cover letters, and personal statements, along with other useful resources to help students navigate the application process.
Let’s connect
I’m interested in academic research positions and industry roles in machine learning, medical imaging, and biomedical data science. Get in touch about opportunities or research collaborations.