For founders building a real product without a senior engineer.
You can write code, or you are willing to learn properly. You are not looking for someone to build it for you.
Build it like you had a CTO. In 8 weeks, or you don't pay.
The Founding Engineer. 8 weeks, 5 founders, one senior engineer in the room. Your architecture decided with you, your code reviewed every week, your product in production by the end. You build it. I make sure it's built right.
5 seats · £2,500 for the founding cohort · in production by week 8 or full refund
What a CTO owns
The job title is the scope. Everything a CTO is responsible for is in. Everything they aren't is out, and I will say so plainly rather than pretend to be useful.
In scope
- Architecture and system design
- Security, auth and data boundaries
- Environments, CI/CD and deploys
- Data modelling and storage choices
- Build versus buy, and vendor calls
- Code review and engineering standards
- Monitoring, alerting and incident response
- Technical due diligence readiness
Not my job
- Sales and marketing
- Your pitch narrative and deck
- Cap table, legals and fundraising strategy
- Design and brand
What you get
One cohort, everything below, one fixed price. No day rate and no hourly meter.
- Your architecture, decided with youNot a lecture on architecture. Your product, your constraints, your call, made with someone who has been wrong enough times to know where it hurts. Written down in week one so every later decision has something to check against.£4,000
- Your agent development system, built properlyCLAUDE.md that actually works, reusable Skills, subagent definitions, Codex wired in as an independent review pass, worktrees for parallel work. Configured against your repo, not demonstrated on a toy one.£2,500
- Weekly review of your real codeOne pull request a week, read properly and answered in writing, for eight weeks. This is the part nobody else sells, because it does not scale and cannot be recorded.£4,800
- Environments, secrets and the deploy pipelineDev and prod actually separated, secrets where they belong, migrations that roll back, CI that catches what you missed. The unglamorous half that decides whether you can ship on a Friday.£3,000
- Technical due diligence packWhat an investor's technical reviewer will ask, and answers you can defend. Most founders get taken apart here because nobody warned them it was coming.£2,000
- Monitoring and the 3am questionErrors surfaced, alerts that mean something, and a written answer to what happens when it breaks while you are asleep.£1,500
- Two sessions, one to oneForty-five minutes in week one to set the architecture, forty-five in week eight to confirm it shipped. The two moments where a group cannot substitute for a conversation.£1,200
- Bonus: every template I useMy own CLAUDE.md, Skills, subagent definitions, model-selection rubric and review configs. Yours to keep and change.£1,000
- Bonus: 90 days of access afterwardsThe group, the recordings and the channel stay open for three months after week eight, because the questions that matter arrive after the deadline.£1,200
A fractional CTO runs roughly £2,400 to £20,000 a month. Lighter-weight CTO advisory, which is closer to this, runs £1,600 to £4,800. Eight weeks of one, alone, is five figures. This is the same person, the same judgment, shared with four other founders. The founding price holds for the first cohort only, because the first cohort is buying something unproven and paying partly in feedback. After that it is £4,000.
How the 8 weeks run
Week 1
Decide
Architecture, stack and the acceptance test that defines shipped. One 1:1 each. Nothing gets built until there is something to build against.
Weeks 2 to 3
Foundations
Environments, secrets, auth, data model, pipeline. The work that is miserable to retrofit and cheap to do first.
Weeks 4 to 6
Build
Your agent workflow at full speed, with your code reviewed every week. Two hot seats a call, so everyone gets their own product on the screen twice.
Weeks 7 to 8
Ship
Production, monitoring, the failure cases. It goes live and the acceptance test either passes or you are refunded.
Exactly how much of me you get
Written down so you can hold me to it, and so nobody discovers the limits in week three. Unlimited access sounds generous and is how these things quietly become worthless for everyone.
- Eight group calls, 90 minutesTwo hot seats each, rotated, so every member is on screen with their own product at least twice.
- Two 1:1 sessionsWeek one and week eight. Not open-ended access.
- One code review a week, eachOne pull request, read properly, answered in writing.
- Questions in the group channelNever by direct message. I answer once and five people learn it.
And the line that decides everything else: I tell you exactly what to change. I don't change it for you. If you would rather it was just done, that is the sprint, and it is a better fit.
The test that decides it
In week one we write down one sentence that defines shipped for your product. Not a scope document. One sentence, in your words, that is either true or false in week 8.
By the end of week eight your product is live on your own domain, with the checks we agreed in week one passing. If it isn't, you get a full refund and you keep everything: the architecture, the configs, the code, all of it. Conditional only on turning up to the calls and submitting the weekly work, because I cannot ship a product you are not building.
After the 8 weeks
- Continuity, after the eight weeks£600/mo
Same cadence, month to month, for the people who want to keep a senior engineer in the room. Offered in week seven, not at the end. £6,000 paid annually, which is two months free. Cancel with a month's notice.
What I ask on the first call
- 01What are you building, and who pays for it?
- 02What exists today, and what is it built on?
- 03Which technical decision are you least sure about?
- 04What happens to the business if this isn't live in eight weeks?
- 05Who else is technical on this, if anyone?
If you are better served by the sprint, or by nothing at all, I will say so on the call. 5 seats is not enough to spend one on the wrong person.
Fit
Works when
- You are building a real product, not exploring an idea.
- You can write code, or you will genuinely learn.
- You can make a decision without a committee.
- You have eight weeks of actual time, not eight weeks of intention.
Doesn't work when
- You want it built for you. That is the sprint, and it is a better fit.
- You want a course on Claude Code. Free ones exist and they are good.
- You need help with sales, fundraising or your deck.
- It is already live and working, in which case you don't need me.
If that isn't you
The cohort is for people building something and doing the building themselves. Most people who land here want one of these instead, and saying so is cheaper for both of us than a call that ends in no.
- Something you already built is broken£4,500
Auth, payments, deploys, data isolation: one launch-blocking problem taken from broken to working in production in three working days. One acceptance test agreed before you pay, and a full refund if it doesn't pass. You don't have to learn anything, and you don't have to be technical.
The Launch-Blocker Sprint → - You just want to know what's actually wrong£350
A written teardown of what is broken, why, and what it takes to fix, inside 48 hours. Credited in full against a sprint booked within 30 days, so it costs nothing if you go ahead.
The Blocker Diagnosis → - You run a business and something is done by hand£7,500
Not a startup problem. Two or three people moving information between systems all week. I find the job that eats the most hours, build the thing that does it instead, and prove it runs on your real data for five days before you pay.
The Bottleneck Sprint →
Why me
There is no shortage of people teaching AI coding tools right now. Most of them are teaching the tool. The tool is the easy part, and the free courses cover it well.
The hard part is knowing which of the six things the model just suggested will still be standing in a year. I was the first engineer hired at Luna.ai, where I took the backend from three people to twelve and past 50,000 sends a day across a thousand customers before Zapier acquired it. I have since built multi-agent systems with real evaluation harnesses, and autonomous navigation software where being wrong has consequences. You are not buying a curriculum. You are buying the judgment.
“Louis was enthusiastic and exceptionally knowledgeable from the start, bridging our knowledge gap, understanding our vision and converting it into reality at speed.”
That was a build engagement, where I wrote the code. No cohort has run yet, so there is no cohort testimonial to show you and I am not going to dress that one up as one.
Written up in more detail
5 seats · founding cohort
Leave your email and I'll send you the start date and the full week-by-week outline before it opens. No obligation, and the waitlist gets first refusal on the £2,500 founding price.
Would rather just talk? Book a free 15-minute call and I'll tell you honestly whether this or the sprint is the better fit.