AI / ML · Clinical Data · Research Software
Alistair Walsh
I build AI-enabled health data systems: clinical pipelines, machine learning workflows, and human-centred tools that turn messy research and care data into usable evidence.
About
I am a Melbourne-based data scientist and research software builder working across clinical data, machine learning, and AI-assisted workflows. My PhD research focused on using artificial intelligence and machine learning to predict recovery trajectories following stroke.
I have built REDCap automations, FHIR/EMR data pipelines, neuroimaging and clinical machine-learning workflows, and researcher-facing tools that make complex datasets easier to query, validate, and explain.
Alongside research and engineering work, I have taught Python, data science, machine learning, Unix, high-performance computing, and digital research skills through universities, The Carpentries, Intersect, and General Assembly.
Focus areas
- Clinical data engineering and interoperability
- Machine learning for health and neuroscience
- LLM-assisted document extraction and review
- Technical training and research enablement
Projects
Cerner FHIR Pipeline
End-to-end data pipeline for extracting, transforming, and serving clinical data via the FHIR standard from Cerner EHR systems.
View project →REDCap AI Ingestion Pipeline
AI-assisted clinical document ingestion using Claude Vision and FastAPI, extracting structured data from unstructured medical records into REDCap with human review.
Case study in progressStroke Recovery ML
Published machine-learning workflows combining neuroimaging and clinical data to model individual differences in recovery following stroke.
View project →Agentic AI Systems Lab
A working lab of AI agents and automation tools for coding, research assistance, monitoring, and multi-step knowledge workflows across local machines.
Case study in progressEvidence
Published stroke recovery ML paper
AutoML and functional brain network patterns associated with post-stroke somatosensory recovery.
Read more →Frontiers in Human Neuroscience coauthor
Optokinetic nystagmus and multistable rivalry between overlapping motion stimuli.
Read more →General Assembly instructor profile
Public teaching profile covering ML/AI for stroke recovery and DSTO physiological cognitive-load work.
Read more →Library Data Carpentry speaker bio
Research training, BCI, neuroimaging, machine learning, and digital research skills background.
Read more →