I help nonprofits and small operators make complex work easier to run, explain, and improve.
Former Chief Program Officer. Creator of PHX Good.
Experience across service delivery, evaluation, logistics, product, and team leadership.
Product direction, evidence rules, and public interfaces—built through a human-led, AI-assisted workflow.
Start with a need, choose a place, and inspect where help begins—including placed access points, source limits, scheduled transit estimates, and an opt-in 3D Valley.
My role: I set the product direction and evidence rules, made the final decisions, and worked with AI coding agents on implementation and testing.
The PHX Good Brain makes public-source relationships among government, funders, organizations, and programs explorable.
My role: I directed the research question and evidence limits; AI research and coding workflows helped assemble the preview.
A public control room separates what is planned, what actually ran, what is blocked, and what remains unknown.
My role: I defined the operating questions and accountability standard, then directed AI agents in building and testing the public record.
The supervisor, the case manager, the director, the funder. Explore what the same number means from each seat.
Interactive essays, field notes, and poetry. Different ways of making an idea land.
Read the writing ↗Move between open book and black box. Notice what changes.
Read the published essay ↗Directions I’m exploring. Each starts with a practical need.
My contribution, the supporting work, and the limits of each example.
7 projects
Case files describe my work and its reported status. Project figures are snapshots from the source portfolio; follow the project links for current records. Sample models illustrate methods and do not establish real-world outcomes.
A short illustrated walkthrough of PHX Good: find a program, check the details, and prepare a referral brief. The case manager makes the contact.
A concise product story, animated interface, and clear explanation of the workflow. The screens are illustrative; the film does not document a real referral or a service outcome.
A concept gets easier to understand when you can see it unfold.
Watch the language change as transparency gives way to concealment. Then take control of the slider yourself.
Explore it yourself ↗Decisions, assumptions, and corrections are available to inspect. A reader can follow the working and ask a better question.
Ready to play · text demonstrationThe September 25 concept kit included a landing-page concept, a market brief, a growth plan, and a 1080p video animatic.
Created as a demonstration of an AI-assisted workflow. This is concept work; no commercial results are claimed.
Operations numbers mean different things depending on where you sit. Each scene here puts you in a chair, at a door, or in front of a board tile, then names the rule.
Sample week, sample families, sample listing. No real clients, programs or organizations. Families are described by facts only.
Each section starts with a rule, puts it in a sentence you can change, and shows the result. Drag any number, or tap it and use the arrow keys. Every model shows its math.
Sample organization, sample numbers. Nothing here is any real organization's data. The point is the method: if you can move the inputs and the result behaves the way the rule says, the assumptions can be explored. Every assumption is listed under "Show the math."
Every package starts with a 30-minute recon call, runs on a model from the analytical examples or a build I've shipped, and names what you'll have by day 10.
Pick the one that sounds most like your week.
Explore the managed digital studio. Choose the parts you need, see how the work would run, and inspect a sample client ledger. Build a sample engagement ↗
I write to make operations make sense to people outside them. Some of it is a new format.
So why is it so hard to check anything? Explore the full new edition, with Thoreau’s itemized receipt, the “Receipt or vibe?” quiz, and an interactive disclosure dial.
Read the interactive essay →Writing, research, and interaction design · AI-assisted
Headings stay fixed while the text moves between two poles, so the reader feels the difference. A short sample in the style of my first published one, Everyone Is Transparent Now.
Three selected pieces on attention, responsibility, and the practice of writing.
The one who Hastened to the door The one that Painted Castle Dread Into the Likeness of himself No one Constructed like himself Who launched Blockades against himself Whose faded Feeling 'fect the self Who faked an image of the world The one who hastened to the door
I never meant to be the burden bearer, but souls need saving. I'd say it's now: I'd never cheat upon or steal the secret through the door. I held my ear to what they said; I wouldn't listen anymore. I knew it wisdom, but I left and told the stories told before. But then I listened, Rose, I couldn't hear their voices anymore.
I write this poetry to document the fever I have seen. The worlds that could have been, what should have been, what dreams conceive to me. They say, 'You shouldn't rhyme as much; the words begin to lose their touch.' These worlds were never meant for them. These words were never meant for them.
A new edition with an itemized receipt, a six-question quiz, and a disclosure dial. Read the full piece →
For case managers: what changed at the door, job fairs, who is hiring, and felony-friendly employers.
What a referral log shows about need and gaps in the Valley, written for reporters and funders.
Plain-language breakdowns of a nonprofit's public filings: where the money goes and what the numbers can and can't show.
Results, year-end reviews, and the mission moment that makes the data land.
Sentences with numbers you can drag, in the tradition of explorable explanations. These examples demonstrate the analytical approach.
I'm an operations leader and builder based in the Phoenix area. My experience at UMOM New Day Centers spans case management, workforce development, evaluation and impact, and executive program leadership across shelter, housing, workforce, coordinated entry, and heat relief.
Along the way I ran workforce development, built an evaluation and impact function, rebuilt a donation center and supply line, and wrote launch plans for new programs. The habit is always the same: make the math visible, then build the routine that lets a team act on it.
Now I'm building PHX Good, a referral tool for case managers, using AI-assisted workflows with human review. I'm interested in roles where operational experience, analytical thinking, and practical product development can help a team deliver better work.
I've served in the Army since 2012: Airborne infantry, the opposing force at the Joint Readiness Training Center, then the National Guard, including a 2023 to 2024 deployment to Jordan. I'm a Sergeant in the Arizona Army National Guard and still teach battle drills and weapons systems.
"A little clarity can change the next step."
Three questions run everything I build. Is it useful and clear? What's in the way, and am I curious about it? Who owns the decision, and what we don't know yet? The answer to the last one is me.
Systems I designed and built myself, each with its honest status. I am the Founder of PHX Good.
An interactive family-history archive combining archival research, maps and storytelling, with evidence and unresolved relationships clearly distinguished.
Concept and creative direction: Fernandez Bjork. Research and implementation developed with AI assistance.
Explore Cusseta RoadA resource site with source-checked guides and a case manager newsletter on the way, so case managers spend less time hunting for current local help. I run the link checks, the phone-width accessibility reviews and the publishing routine.
phxgood.org ↗A written shelter policy turned into a decision framework, then into practice scenarios with AI roleplay and coaching feedback. Built as a training tool for staff who need to apply a policy the same way under stress.
One app with a private workspace for each client. A verified email code is the front door, the server matches it to a single workspace, and nightly backups are planned. Next up is proving that one client can never see another.
Mobile-first sites and app-like portals for a local church, a service business and a community group. Each gets a tile home, a questions card that counts open items, and a launch checklist that ends with the client's written approval.
A private app with a swipe deck of matched roles, scoring, reviewed letters, and duplicate and eligibility checks with sources. Nothing is sent without approval, and sending is off by default.
Scripted jobs on a local model that read scanned records, sort files and cross-check results against a master log. Every job has a check, and nothing leaves my machine.
I can hold a Power BI model and Whitman in the same head. That combination is most of what I bring.
I take blame downward and give credit outward. When something breaks on my watch, I say so first.
I see the second-order problem. Occupancy isn't a number, it's a funder relationship and a family waiting.
I walk in with the numbers, the names and a written plan. I bring receipts to every room.
I make strangers feel like people, and I build teams that show up for each other.
I name my gap before someone else has to, and I don't inflate a number to win a meeting.
DiSC: Motivator (high I, moderate S and D)
