Matt Joyce

National Manager, Emerging Technology and Health Informatics, Calvary Health Care

This is an identity artifact: a place where my work, thinking, and projects stay coherent over time. It exists so that I am legible to both humans and the algorithms that represent us.

Last updated: 2026-02-20

I'm a strong match for

Capabilities

Healthcare AI governance and responsible adoption

Methods: capability boundaries and MVP scoping, use-case triage and risk routing, policy-aligned guidelines and guardrails, stakeholder translation (clinical, legal, exec, cyber)

Constraints: regulated environment, clinical safety adjacency, privacy and PHI sensitivity, enterprise change control realities

Outcomes: clear decision pathways for acceptable vs controlled AI use; reduced shadow-AI risk through usable guidance; stakeholder-aligned framing for governance decisions

Automation program shaping (enterprise)

Methods: roadmap creation and intake/rubric design, workflow-to-system mapping, practical governance that enables delivery

Constraints: multi-stakeholder environment, operational variability across sites, limited delivery capacity

Outcomes: actionable automation backlog structure; prioritisation logic that leaders can defend; MVP definitions that can be executed

Knowledge systems / corporate memory (RAG + graph-shaped thinking)

Methods: schema-first thinking for traceability, entity + relationship modelling, retrieval and evidence-oriented representation, lightweight architecture decisions to avoid overbuild

Constraints: long-term maintainability, token/compute cost awareness, traceability requirements

Outcomes: structured representation of decisions and context; improved retrieval of institutional knowledge; better reuse of prior work through evidence objects

Pragmatic prototyping (internal tools / personal APIs)

Methods: Python scripting, FastAPI services, SQLite-backed services, YAML/JSON schema patterns, operational automation (cron/systemd style)

Constraints: security hygiene, keep it small and evolvable, avoid toolbox-fallacy overengineering

Outcomes: working prototypes suitable for iteration; clear separation of concerns and config; deployable patterns for small services

Boundaries

I do: governance and capability boundary design, workflow and product discovery framing, pragmatic prototyping guidance, system design reviews

I don't: share or handle PHI in unsecured tools, work that depends on deception or dark patterns, 24/7 operational responsibility commitments

Conditions: doc-first / async preferred where possible, explicit constraints and decision context, clear owner for decisions (avoid analysis-without-ownership)

Engagement

Evidence