Choose
Find a problem worth solving: five user conversations, a value hypothesis, and a written kill-or-commit decision.
Six-week education and coaching cohort
Bring one real problem and take it through discovery, scope, build, verification, release, and handoff, with a written review from a senior software engineer at every gate.
November round · enrollment closes October 15 · $499 paid in full · 6–8 hours a week · capped at 10 builders

Bring one real problem. Leave with a reviewed release.
Read this before you join
6weeks
10builders max
3written reviews
1real release
The backstory
One founder I work with.
The automations. He automated his own operation for years, one flow at a time, each one solving that week’s problem. Together they hardened into a machine nobody could safely touch: change one piece and three others broke. By the time I met him the technical debt had the whole business strangled, and I have been dismantling it with him piece by piece, replacing it with a system his team runs on the floor every day.
He had the logic of his business completely worked out before I met him. What was missing was the built thing. Capstone exists for the founder at that exact point.
The Capstone delivery system
AI made prototypes easy. The hard part is everything between a promising demo and something another person can trust and use. That is what I teach.
Find a problem worth solving: five user conversations, a value hypothesis, and a written kill-or-commit decision.
Reduce it to a valuable six-week vertical slice, with acceptance criteria, a backlog, and explicit non-goals.
Use AI to write the code without handing it your technical judgment: version history, tests, and safe data handling.
Verify the workflow with realistic data and watch at least three real people use it.
Deploy it, or complete a real handoff to your stakeholder.
Ask for payment, adoption, approval, or another concrete commitment. A documented “no” counts as evidence.
Six weekly gates
Every week: one 75–90 minute education and build lab, one 60-minute office-hours clinic, a short pod check-in, and one visible milestone. The rest is your own build time. Sessions are recorded.
Get ready. Supported stack, repository basics, data and cost constraints. Readiness clinic and pod matching.
Choose the problem. Five conversations or observations, a one-page brief, and a kill-or-commit decision. Review 1: problem and scope.
Cut scope. Story map, Definition of Done, prioritized backlog, decision log, and your first runnable slice. Live scope teardown.
Build and verify. The core flow on realistic data, with verification evidence and an updated risk register. Review 2: working slice.
Release to reality. A live environment, three observed user sessions, and an issue log. Release clinic and missed-milestone rescue.
Test value. A priced offer, pilot request, adoption request, or before-and-after measure. Review 3: release, value test, and handoff.
Deliver and hand off. Evidence pack, handoff README, demo, retrospective, and a thirty-day plan. Demo Day.
Weeks one and two go to problem choice and scope on purpose. That is the skill that stops a project stalling at 80 percent.
The graduation standard
You graduate on evidence, not attendance. Graduation does not require revenue. It does require proof from outside the cohort.
Client work
What a client says about the delivery work behind this cohort.
“They exceeded all expectations. They are true experts in this field, delivering high-quality work with precision and efficiency.”
This quote is from client delivery work, not from a cohort graduate: cohort one has not run yet, and its results will be here when they exist.
Education and coaching
I teach the delivery method, then coach you through applying it to your project. You make the decisions and build the product. I review your scope, challenge your assumptions, inspect the evidence, and help you get unstuck.
That means three written reviews at the gates that matter, weekly office hours, and an async blocker channel with a stated response window. It does not mean unlimited one-to-one time, me taking over your repository, custom integrations, or a production warranty.
This is education and coaching. I do not build the product for you, promise a job, promise customers, or promise income.
And six weeks will not turn you into an engineer. It will take one real project out of the demo stage and teach you the way of working that got it there.
Where Capstone fits
Knowledge and community stay free. The cohort sells deadlines, review, and accountability. The partnership is for problems that need me hands-on.
Starter lessons, templates, demos, community Q&A, coworking, and challenges. Explore, find a problem, meet peers. No individual review or graduation. Join the free community →
One standardized delivery cycle with practitioner feedback: six-week curriculum, pods, build lab, office hours, three written gate reviews, Demo Day, and evidence-based graduation. You build. Join the cohort →
For consequential, context-specific production problems: a private roadmap, custom architecture, stakeholder facilitation, and hands-on implementation. Selective and individually scoped. Ask about it →
Founding cohort
November round · Enrollment closes October 15 · Capped at 10 builders · Live and recorded
$499
Paid in full. Enrollment and payment happen on Cohort Camp, where the cohort’s lessons, pods, and Demo Day live.
Full refund through the end of week one, once you have attended onboarding, submitted your problem brief, and completed the readiness conversation. After that, you can transfer once to the next cohort.
The round runs if at least 8 people join. If it doesn’t fill by the deadline, it’s cancelled and you choose: a full refund or a place in the next round.
Join the founding cohortThird-party tool and API costs are separate and disclosed during the readiness week.
Your cohort lead
David Yu has more than a decade of senior engineering experience. At Futureman Labs he builds internal tools, automations, and client systems that have to hold up after the demo.
You are getting his real delivery process: narrow the problem, build the smallest useful version, test it in context, and keep going until people use it.
Q&A
No professional coding background is required. Expect technical setup, debugging, and guided work in development tools. You drive the build with AI; I teach and review.
It is education plus coaching. There is a curriculum and every session is recorded, but the point is applying the method to one real project and graduating on evidence, not watching videos.
That is fine, as long as you have a specific problem and can reach at least five prospective users or one real stakeholder. Weeks one and two are about choosing and cutting the problem.
A small product or micro-SaaS, a fixed-scope client pilot, an internal workflow or operational tool, or a portfolio or public-good product with real external users. One project per participant.
Plan for 6 to 8 focused hours: the build lab, the office-hours clinic, a short pod check-in, and your own build time.
Maybe, but I do not promise it. Every participant makes one ethical offer or asks for a concrete commitment. Revenue is useful evidence, never a requirement to graduate, and never guaranteed.
You keep the release, the evidence pack, and a thirty-day plan, plus continued access to the free Cohort Camp community. If your next milestone needs custom architecture or hands-on work, that is what the Implementation Partnership is for.
Yes, in full through the end of the first week, once you have attended onboarding, submitted your problem brief, and completed the readiness conversation. After that, you can transfer once to the next cohort.
Last call
Bring one real problem. Leave with a reviewed release and the evidence behind it.
Join the founding cohort