New: the Pi Worker Model, my patent-pending framework for human-AI teams. Try the interactive model
Product · Org design · AI
Operator, founder and human-centered product leader. I helped scale General Assembly's engineering programs from 1 to 12 markets on its way to a $412.5 million acquisition, and I've done similar work from startups to larger tech companies. Now I build AI that people trust and communities benefit from.
I advise, lead and build with foundations, public agencies, nonprofits and AI companies that want AI to work for people.
Why I do this
Most AI efforts don't fail on technology. They fail on trust, adoption and change. That is why I connect product, organizational design and AI.
The question I ask of every system I build: does this expand a person's capacity to think, or does it train them to require less thinking of themselves?
What I stand for
The people closest to the work know where it breaks. At Open Door Legal, staff could ask anything about AI anonymously, and what they asked shaped our position.
Mandates get compliance. Choice gets adoption. I design for the yes people can give freely.
You can't bolt it on at launch. It lives in audit trails, in human review when confidence drops, in rules checked while the work happens.
Poetry and breathwork taught me pacing. Systems need it too: a rhythm people can rely on, and a pause when something matters.
Taking General Assembly from 1 market to 12, I learned the curriculum travels easily. Culture has to be carried.
Ways to work together
Advising
You've decided to bring AI in. I help you decide how, so your people trust it, use it and stay in charge of it.
State Bar of California AI Technology Champion. Built an AI readiness model to help adoption scale across legal aid organizations.
Talk about advising →Leading
I take the work from idea to everyday use: the product, the people who run it, and the change in between.
17+ years leading products and programs, from Dev Bootcamp and General Assembly to Open Door Legal.
Talk about leading →Building
We start where the work actually happens, with the people who do it, and build something small enough to learn from and real enough to matter.
Led Open Door Legal into Salesforce's Agents of Impact Accelerator, which accepted under 5% of applicants.
Talk about building →Speaking and workshops
Keynotes, panels and workshops on the human side of AI. Sometimes I bring data. Sometimes I bring a poem.
Featured speaker on Salesforce's Agents for Impact webinar. Led an AI adoption workshop series in summer 2026.
Invite me to speak →17+
Years building products and organizations
1 → 12
Markets for General Assembly's flagship engineering programs, scaled as an early team member
<5%
Acceptance rate of Salesforce's Agents of Impact Accelerator, which I led Open Door Legal into in 2025
$1.5M
And 1,000+ students, from three products I launched as Dev Bootcamp's first product hire
$4M → $10M
Open Door Legal's growth as a nonprofit since I joined, now the largest legal aid organization in San Francisco
1
Patent-pending framework for how humans and AI agents collaborate
What I bring
I help leaders bring AI into their organizations in a way people trust, understand and choose.
Designing roles, handoffs and trust between people and AI.
Where does the agent stop and the person start? I make that line visible, then design the handoffs so nothing important falls through.
From discovery to launch, with the frontline in the room.
Discovery, prototypes and pilots shaped by the people who will rely on them, then launched with training and a feedback loop that keeps improving the agent.
Policy, guardrails and accountability people can actually follow.
Plain-language policy, staff training and earned access, grounded in California's professional conduct rules and the State Bar's guidance on generative and agentic AI.
Adoption that sticks because people chose it.
Champions, rituals and feedback loops, never just a memo. I design the change plan alongside the product, not after it.
Patent pending
Most organizations still treat AI as a tool on the side, and most people were never trained in prompt engineering. The Pi Worker Model turns AI into a teammate inside a governed system: a shared rhythm, new hybrid roles, agents that translate what people mean, and autonomy that adjusts to the stakes.
People and agents align on priorities and context.
Agents carry steady work. People own judgment and care.
People review and coach. Every correction teaches the system.
Like sorting an inbox or drafting a meeting summary.
How sure the agent is about this particular piece of work.
The agent drafts. A person approves.
Routine steps move on their own, and nothing reaches a client until someone signs off.
Confidence is healthy. Review happens in the evening feedback, and every correction trains the system.
Bringing AI into your organization? Let's talk about the human side.
Governance people can follow
Rule one: AI is always optional.
The first line of the AI Use Policy I wrote for Open Door Legal. Then I turned the policy into practice.
Three golden rules: protect sensitive information, verify everything, be transparent.
A staff series built around a TRUST checklist for verifying AI output.
Tools come after people read the policy, train and sign an acknowledgment.
An anonymous channel reviewed monthly, and open sessions where staff shape the position.
Selected for the State Bar of California's AI Technology Champions committee, bringing the frontline view from legal aid.
As a product board member at TrustBound, I build governance and regulation products for humans, agents and AI tools.
Wrote Open Door Legal's security incident response policy to meet a city funder's requirements.
Built an AI readiness model so adoption can scale across legal aid organizations, not just one.
Journey
B.A. in Anthropology, Biology and Economics
Anthropology taught me to watch before I fix. Most problems are rituals before they are processes.
First product hire. Launched three products
Students quit jobs to bet on what we built. Every feature was a promise with someone's rent attached.
Early team. Scaled engineering programs from 1 to 12 markets. Acquired by Adecco Group for $412.5 million
Twelve markets in, I learned that whatever you tolerate early becomes what you scale.
Head of Product
In health, the person waiting on your product is already tired. Speed is a form of care, and so is getting it right.
Senior Program Director, with 150% enterprise growth
The technology was never the slow part. People were, and they were usually right to be.
Founder, organizational design firm
Every org chart is a theory of who gets to decide. I started redesigning the theory.
Director of Innovation and Strategy · State Bar of California AI Technology Champion
A client's story arrives uneven and afraid. Good technology lets it stay human long enough to be heard.
Product board member, building governance products for an agentic workforce
Trust is infrastructure. Now I build it.
From my writing
What concerns me, finally, is not that machines will think for us. It is that thinking itself may become optional, and then rare.
A woman facing eviction types her story into a text box, and a system quietly teaches her how to be legible. I help build tools like that, and I believe in them. This essay is about what they cost: who still gets the time and quiet to think, and the question I now ask of every AI system I touch.
The whole person
I'm a poet, spoken word performer and breathwork facilitator. The attention I bring to a line of poetry, or to a room breathing together, is the attention I bring to the systems I design.
Breathe with me. One breath.
Let's build it together
I love hearing from mission-driven teams, foundations, public agencies, civic tech and AI companies, whether you're shaping a strategy, building agents or writing the rules.
Or write directly: [email protected]