Senate → Gates → Wall Street → AI
I built an enterprise AI function from scratch in one of the hardest rooms there is — regulated, skeptical, high-stakes. That playbook travels.
The Pattern
A whole firm. No mandate. Every dot is a choice.
Governance built first and put in front of leadership. Every sign-up after this is a choice.
Case Studies
Real examples of what I've built, led, and transformed.
Speaking
Where I've shared what building enterprise AI adoption from the ground up actually looks like, from conference panels to the training room.
Panel · Sept 2026
How to drive AI adoption without mandating it: tie every training to a real business context, never a tool demo, and pace it deliberately. Sticky beats fast.
AI doesn't replace a process, it exposes where the old one was already broken.Read my recap on LinkedIn →
Podcast · Coming soon
Why I never mandated use, why enthusiasm isn't adoption, building inside legal and compliance's own workflows first, and why I stopped building tools and started teaching people to build their own.
The real test: can people build and get value without me in the room?See the episode preview on LinkedIn →
Training
From firmwide sessions to group-specific deep dives, every training is tied to a real business context, never a tool demo, and paced so the habits stick.
Interested in having me speak? Get in touch →
The Live Brain
I run a second brain that mines my own work and distills the rules I operate by. The method is open-sourced in the Workshop below. Open any rule to see it in practice, and where it's running on this page.
I ask my own AI output point-blank: did you make anything up? It has said yes. Those lines die on the spot.
Where a number can’t be shared, it isn’t rounded or hinted — it’s removed. That’s this rule running.
My automations run under standing orders: if the week’s results are thin, say “thin.” Padding a list to look busy is a firing offense.
Every skill the Workshop below builds carries the same no-padding rule.
In anything I ship, sourced numbers and strategic opinion are visibly different things — a reader always knows which one they’re holding.
The facts on this page are the verifiable kind. The opinions live in the essays, signed as mine.
When someone asks whether AI can really run a workflow, they get a working tool in their hands by Friday — not a deck about one.
This site. The skill generator below. infra-loop on GitHub.
Before restructuring anything — a pricing model, a rollout, a migration — I pull the real usage data. Assumptions don’t survive contact with actuals.
The pattern above is drawn from the method, not the metrics — the real numbers stay inside the firm.
Models I ship get deliberately perturbed before they’re trusted — a self-test that has to fail correctly, or the work isn’t done.
Every skill the Workshop builds ships with a self-check step for exactly this reason.
No tool reaches users before the policy, the training, and an acknowledgment gate are standing. Access is earned, not defaulted.
The Governance & Trust case study below — guardrails as the accelerant.
Each caught error becomes a permanent verification rule. The rulebook is a library of scar tissue — and the reason leadership trusts the output.
Those verification standards are the spine of the case study below.
A one-week dip once read as decline; the trend showed a weekday cycle. I report trends and disclose definitions — snapshots flatter.
It’s why the timeline scrubs across a year instead of flashing one number.
Before any big ask, I write the skeptic’s brief against my own case and argue it down first.
It’s why the pitch generator is allowed to hand out a low fit score.
Tool choices ship in days and get corrected in hours. Anything irreversible gets staged, deliberate, multi-pass review.
This site iterates weekly. The claims on it don’t move without proof.
Rollouts start inside the real workflow: friction removed first, fears answered before they’re voiced, leadership visibly going first.
Voluntary is the only adoption I trust — nothing on this page was mandated either.
Training never opens with features — it opens with the person’s own work. Once they’re fluent, they design the next tool for me.
The Workshop at the end of this page exists to hand you that same fluency.
The values underneath. Tap one to see the rule it turns into.
distilled from a real system · people and specifics de-identified by design
Methodology
The playbook behind the results: three moves, on a loop.
Find where people really are.
Before building anything, understand the real problem. Map stakeholders, identify friction, and find where people are today — not where you assume they are. Then I map the system — every workflow, dependency, and handoff — so I know exactly what to re-architect for AI.
The actual thing, not a deck.
I build the tools myself — dashboards, MCP integrations, Skills libraries, custom apps, training content — then iterate based on what real usage data tells me. Not decks about what to build. The actual thing.
Built for the humans who use it.
Technology without adoption is just an expense. Design every rollout for the humans who'll use it. I re-architect the workflows, write the training materials, build the scoring systems, and stand up the governance. The architect and the general contractor.
The Workshop
A whole library of them runs in my day job. These three are yours — make one, take one, or read the method.
Describe something you do at work every week. Answer three questions. Leave with a working skill — my verification standards included.
The skill that built this site — resume in, deployed site out, in one sitting (about an hour). Netlify, Vercel, or GitHub Pages.
Download .skillinfra-loop: a second brain that mines your own work and writes its own skills, on a loop. Free to share and adapt.
View on GitHubKeep it generic — no confidential names, clients, or data. Your answers aren't saved; the only thing counted is that a skill got built.
The site skill requires the Claude desktop app with Cowork mode enabled
Career Journey
Seven chapters, four worlds, one toolkit, and every piece of it shows up in the AI work.
U.S. Senate, Office of Senator Dean Heller
I ran the operations of a Senate office: the budget, payroll and compliance, a team of four, and coordination across five committees. It's where I learned to operate when every detail is public.
Pivotal Ventures
I turned a principal's time and priorities into execution: logistics, relationships, and the process work that keeps a fast-moving portfolio on track.
Gilder Gagnon Howe & Co.
My first seat in finance: client strategy, board prep, and executive operations for a Managing Partner.
Geller & Company
As strategic partner to the CHCO, I led HR and technology workstreams through the Bloomberg acquisition, running a cross-functional PMO end to end.
Summit Trail Advisors
I built a Central Planning function and redesigned pricing and process as the team scaled.
Perella Weinberg Partners
Data analytics, platform buildout, and budget management across the firm, the systems view that set up everything next.
Perella Weinberg Partners
Everything converges. I built the firm's enterprise AI function from zero to firmwide, voluntary adoption across 12 global offices: governance first, net-neutral funding, and training tied to real work.
About Me
Pick your depth.
I build the bridge between what AI can do and what organizations will actually adopt, by re-architecting workflows and building the tools myself.
Reno, Nevada. One of five kids in a blended family, working since 15, and I put myself through college at UNR.
I don't hand off specs. I build the dashboards, write the training, design the MCP integrations, and stand up the governance myself.
The moment someone goes from skeptical to "wait, this changes everything." I've watched it hundreds of times, and it never gets old.
The parks with Farley, my 80-pound goldendoodle, plus reading, friends, and browsing upstate cottages with a bit of land.
I grew up in Reno, Nevada — one of five kids in a blended family, surrounded by the Sierra Nevada, Lake Tahoe, and the kind of outdoors that teaches you to figure things out on your own. I played volleyball for years — high school and a club team in Northern California — before trading the court for a work schedule that started at 15: first a paint-your-own pottery studio, then a preschool classroom to pay my way through college at UNR — where I majored in Political Science and minored in Cultural Anthropology — with a stint as a certified ski and snowboard technician up at Tahoe mixed in.
In college, I interned for a U.S. Senator while moonlighting as a paralegal and private bookkeeper for litigators and individuals navigating the healthcare system. I presented research on "The Effect of Institutional Characteristics of Public Support for National Legislatures" at the Western Political Science Association conference in Hollywood. I was also in a sorority. The range started early.
That through-line — learning fast, translating between worlds, doing the work nobody else wants to do — carried me from the Senate floor to Melinda French Gates' office to Wall Street to leading an enterprise AI transformation. I don't hand off specs to a dev team — I build the dashboards, write the training content, design the MCP integrations, and stand up the governance frameworks myself. The conviction comes from being in the weeds, not above them. I'm a lifelong learner: every new chapter has been an excuse to go deep on something I didn't know yet, and that curiosity is what makes me good at what I do. What drives me isn't the technology itself. It's the moment someone goes from skeptical to "wait, this changes everything." I've watched that happen hundreds of times, and it never gets old.
Finance has been the proving ground — the adoption problem I've worked on there is the one every high-stakes domain is now facing.
I may live in New York now, but nature is still a huge part of my life, with the best parts of my week spent in the parks with Farley, my 80-pound goldendoodle. When I'm not walking him, I'm reading, researching, spending time with friends, thinking about what the world will look like in five years, and browsing real estate upstate for a little country cottage with a bit of land.
Different readers need different depths. Someone scanning between meetings gets the headline in ten seconds, a curious peer gets the highlights in a minute, and someone who wants to know who's behind the work gets the whole story. It's the same content, layered by how much time you have. That's how I design training and rollouts, too: meet people where they are and let them go deeper when they're ready.
Skill Map
Enterprise AI adoption isn't a technology problem. It's governance, politics, process, and people. Every one of those needs runs back to more than one chapter of my career. Tap either side to trace the lines.
Perspectives
Ideas I keep coming back to — shaped by a decade across government, philanthropy, and finance.
Coming soon
A set of essays on enterprise AI — adoption, governance, and building in high-stakes rooms — is in final edit and lands here shortly. Want a note when they publish? Email me.
If we're in touch about working together, this space is for you. Enter the password I shared with you, or email me to request access.
Interactive
Tell me about your role and I'll show you exactly why.
All views and opinions expressed are my own and do not represent the views of any current or former employer.
Fill in your name and company, then watch the magic happen.
Let's Connect
Whether you're wrestling with enterprise AI adoption, bringing AI into a high-stakes domain for the first time, or building these tools and curious how they land on the customer's side of the table — I'd love to compare notes.
What brings you here?
or find me at
For nonprofits: I'm passionate about helping mission-driven organizations leverage AI. Just reach out directly. I'd love to help.