Operations transformation
Keeping a multi-site humanitarian operation coordinated
The problem: rapidly changing arrivals, multiple shelter sites, hundreds of staff and volunteers, transportation needs, service coordination, and public reporting.
What I built: practical workflows and low-cost data systems using the Google ecosystem to track arrivals, departures, demographics, travel needs, capacity, volunteer coordination, and service delivery.
Scale: a $12 million operation serving up to 1,800 people nightly with 62 staff, more than 600 volunteers, and up to 2,500 operational data points per day.
Independent operational context →
Related coverage: AZPM on the 2024 funding cliff · KJZZ on migrant-care funding pressure · KGUN on opening the Benedictine Monastery shelter
Multi-agency implementation
Helping organizations share referrals and data more reliably
The problem: border shelters and destination organizations needed better continuity of information as families moved across the country.
What I worked on: shared referral workflows, common data fields, reporting practices, documentation, and cross-agency communication across Catholic and humanitarian partners.
Scale: coordination across multiple agencies with high-volume migration-service data, including work involving up to 10,000 new records per week.
Decision support
Turning changing field conditions into information leaders could use
The problem: decision-makers needed usable information about shelter capacity, community readiness, transportation, social services, and operational risk.
What I built: dashboards and executive-ready reporting for government officials and humanitarian stakeholders, combining field information with service-capacity analysis.
Why it mattered: the goal was not a better-looking dashboard. It was to make resource planning, referrals, and operational decisions clearer.
AI & automation lab
Experimenting with local-first AI for operational knowledge
The problem: complex workspaces lose time when knowledge is fragmented across files, dashboards, tools, and services.
What I am building: a governed local-first AI environment combining retrieval, local language models, automation, dashboards, tool access, and human approval gates.
Why I care: I am interested in AI when it reduces repetitive work, improves access to information, or helps people make better decisions. I am much less interested in using it simply because it is fashionable.
How I approach systems workI start with the people doing the job, the information they actually need, and the constraints they are working under. Then I build the simplest structure that can reliably support the work and improve over time. See
Leadership, Implementation & Systems capabilities for the broader skill set,
Research & Evidence for the evidence-to-practice perspective, and the
Implementation Resume for a role-focused version of this experience.