Direct AI to build a real app with a working engineer beside you. Six weeks later it's live at a URL you can hand to anyone who asks what you've built.
We teach you to go beyond basic vibe-coding — and build useful, impressive tools that solve a real need.
About half of each cohort hasn't. You'll learn to direct AI, review what it builds, and get something real working.
You'll go deeper. More complex projects, real deployment infrastructure, and harder problems. We won't slow you down.
Most of each week goes to the real work of building: pinning down the problem worth solving, designing the solution, then directing AI to write and review the code. A working teacher and agentic-coding expert sits beside your student to explain what's happening — with a team of domain experts on call when a project needs one. No syntax drills, no leetcode warmups.
Every student ships a deployed app at a public URL and gets coached on how to write and talk about it. For the Common App project section, for portfolios, and for the interview where someone asks what they've built.
Your student's mentor is the founder, a current classroom teacher with production code to his name. Eight seats is the honest capacity of one person doing real 1:1 work, not a number we picked for scarcity.
Most “AI summer programs” end in a folder of notes. This is the only arc we run — and we don't stop the clock until the last step is real.
A tool you wish existed — for your team, your class, the thing that bugs you every day. No code required to walk in.
Working sessions, not lectures. You direct the same AI tools professionals use — Claude, Cursor, v0 — while an engineer who ships real code sits in the room beside you.
Not a slide deck about a tool. The actual thing, deployed somewhere anyone can open it.
If your URL isn't live by demo day, the cohort keeps meeting until it is. That's the whole offer.
Six weeks of live sessions, a dedicated 1:1 mentor, and a deployment coach who stays in the room until your student has a working URL to share with anyone who asks what they have been building.
We are not a college-admissions service. What your student leaves with is a deployed thing they made themselves, and that ends up being a better answer to “tell me about something you built” than a certificate. Demo day is open to family, friends, and an admissions counselor if you'd like to invite one.
Six weeks of live sessions, 1:1 mentorship, and a demo day at the end. Builders aged 14 to 18 leave with a deployed project running at a public URL.
A short application of about 10 minutes with real questions, not busywork. Then a 15-minute call with your student and you. Eight seats and six weeks of 1:1 means fit matters in both directions. You'll have an answer within two days of the call.
A solo-run program with a teenager deserves more than a logo and a promise. Here is exactly how it is set up.
Hi, I'm Chris. I built this for the kid who already has the idea, and just needs someone in the room who can actually help them ship it.
I'm a teacher who also writes real code, so I get both halves of this. I know how to explain a thing until it clicks, and I've shipped the kinds of apps you'll be building. The ones that actually go live and that people use.
I won't talk down to you. If you've never coded, that's normal, about half of every cohort hasn't. You'll learn to direct the AI, check its work, and get something real working. If you already code, we go deeper.
The whole job is to get you to “done.” Plenty of people can hand you a tutorial. The hard part is the messy middle, where most projects quietly die. That's exactly where I sit next to you until your thing is live.
For the families reading this: about 15 years in education, degrees from Brown, LMU, and UCLA, and a stack of tools I've built and run at school. The short version, your student is in experienced hands.
$1,100 is the founding rate, with $1,850 as the standard rate after the first cohort. There are eight seats because this is a solo-run program with a real 1:1 mentor per student for the full six weeks, and eight is the actual capacity rather than a number we picked for scarcity. For that, your student gets twelve live build sessions, a dedicated mentor, deployment coaching until the project is genuinely live, story coaching on how to talk about it, and a demo day. A private CS tutor at the same contact hours runs well past this and would not produce a shipped project at the end.
That's the one outcome the whole structure is built to prevent. We set scope small and realistic in week one, and the mentor's specific job is to make sure one thing ships. A small working tool is the win we are aiming for, and the structure is designed to keep ambitious-but-unfinished projects from being the alternative.
We don't promise admissions outcomes, because nobody honest can. What we do promise is a specific, public project your student can talk about, link to, and keep improving after the cohort ends, along with coaching on how to write and talk about it. That ends up being a defensible answer when an admissions reader or interviewer asks them to describe something they have built, in a way that generic activities and test prep usually are not.
No. Students with coding experience work on more ambitious projects — production databases, real auth flows, multi-page apps — and spend more time reviewing and directing the AI rather than just prompting it. Students without a coding background learn how to direct AI, check its output, and ship something real. About half of each cohort starts with no programming background. We care more about curiosity and follow-through than prior syntax.
Not if they want to ship something real. Knowing syntax is different from scoping a project, managing an AI pair-programmer, handling deployment, and finishing. Strong coders move faster and tackle harder problems — multi-user apps, real databases, external APIs — but "something deployed and working" is harder than it sounds regardless of starting experience. If your student can already ship independently, this isn't the right program. If they have skills but have never actually finished and deployed something of their own, it probably is.
If your student takes AP Computer Science Principles, the project they build here can usually be submitted as the Create Performance Task. Just ask and we will help line it up. It is a bonus, not the point of the program.
Very little data, and you stay in control of it. We collect what it takes to run the program and never sell or share any of it. Session recordings exist for your transparency, are available on request, and are deleted a year after the cohort. During the build, your student opens accounts on the real tools (GitHub, a hosting platform, an AI assistant) with you in the loop on each one, and those accounts, the code, and the deployed app belong to your family, not to us.
No slides, no pitch. A real tool, built with AI and deployed to a public URL, with the boring parts cut out. The film drops before the founding cohort starts.