Dublin · Medical AI & research engineering
I work across machine learning, biomedical data, intelligent sensing and the practical systems that make research reproducible and useful. I care about the path from a defensible question to an inspectable system—not only the final model score.
Each figure is explained and bounded in its case study.
Selected work / 01—05
Questions, methods, reported evidence and limits—presented together so the reader can judge the work rather than decode a list of tools.
View all projectsA learned image codec that combines a scale hyperprior with context modelling, semantic gating and attention to preserve structure at low bitrates.
A retrieval-first assistant for Irish clinical and biomedical guidance, built around traceable ingestion, hybrid search and exact chunk citations.
Residue-level ligand-binding-site experiments combining ESM-2 protein-language embeddings with graph attention over protein structure.
An offline, interpretable comparison of classical and deep models for three-state stress classification from engineered EEG features.
The tested audio foundation for an offline multilingual subtitle system—carefully separating what works now from the larger product still to be built.
How I work
I want to know what a system should refuse, where a metric misleads, and which assumption will break first.
Manifests, exact citations, explicit splits and narrow status labels make a project easier to question—and stronger for it.
Models eventually meet APIs, sensors, latency, hardware and people. I enjoy working in that seam.
About the trajectory
My undergraduate work moved between hardware, web systems and machine learning. The recurring interest was never a single stack; it was how information moves through a system, where it degrades and how decisions can remain intelligible.
That led from crop-surveillance prototypes and neural compression into protein graphs, EEG analysis and evidence-grounded clinical retrieval—and now to postgraduate medical-AI study at UCD.
Read the full storyResearch conversations
If you’re working on a biomedical ML problem, evidence-grounded system or sensing challenge, I’d be glad to compare notes.
Write to me