We start where courses end.
No beginner hand-holding. Two weeks, Monday to Friday, 9am to 3pm. Production tools including Claude Code, GitHub, Vercel, Supabase, Composio, Playwright, and more. Updated before every cohort.
Four tasks. About an hour.
The rule is gather, don't build — so Day 1 starts at full speed.
Set up your accounts
Claude, GitHub, Vercel, Supabase, Warp, and the Anthropic Console (with $10 in API credits). We send exact links.
Have your first expert conversation
Talk to Claude about your field like an expert would. Save the conversation — you'll use it on Day 1.
Know your problem
Four short written answers: what the problem is, who it affects, what inputs it needs, and what "solved" looks like. This is also our fit check.
Bring your Materials Kit
A real dataset, a few documents, and a list of about 10 real inputs (redacted). Every hands-on day runs on your own material, not toy data.
Day by day.
Product by product.
Set up the tools. Ship Product 1 by Friday.
How LLMs actually work. The real differences between Claude, GPT, Gemini, and DeepSeek. Talking to Claude like an expert, not a search engine.
Warp Terminal, first GitHub push. Draft your capstone PRD with Claude. Scope before you build.
Build a real landing page in plain English. Connect GitHub to Vercel. Push to main, auto-deploy in 30 seconds. Live URL by end of day.
Databases without jargon. Table, form, dashboard, all wired together. Build your first Claude Skill. Learn how AI connects to your tools.
4-minute live demo to your cohort on real data. Then set up a content habit with Claude. The proof-of-work that gets clients and jobs.
Build a real AI agent. Demo Day Friday.
Pipelines vs. agents. Connect Claude to Gmail, Slack, HubSpot, and 200+ tools via Composio — no integration code. Build automations that run on your real data.
Playwright for sites with no API: navigate, extract, schedule. Voice stack: Whisper transcribes, Claude reasons, ElevenLabs speaks.
Claude Code as a full project collaborator in Warp. Ship a working AI agent that connects to an external tool. Deployed on Render, running on your own data.
Learn to test and trust your AI — evals, document Q&A, and handling data and access safely. Full capstone build, deployed by 2:30pm.
Austin tech community in the room. 5 minutes each. Live demo on real data, hard questions, real feedback from builders and investors.
Want the full 2-week curriculum?
Full breakdown of our 2-week in-person program in Austin. We'll also answer any questions about the program.
A site. An agent. A workflow.
All deployed. All yours.
Your first real build
A live site, data tool, or Claude Skill — deployed to a public URL and demoed live to your cohort on Friday of Week 1.
A working AI agent
A real automation or AI agent — connecting Claude to your actual tools via Composio, browser control, or voice, running on your own data.
Your main project
Your most ambitious build — polished, deployed, and presented live on stage at Demo Day in front of Austin's tech community.
Real tools. Not toys.
Every tool in this program is used in production. You'll know when to reach for each one.
Real, working tools. Two weeks.
Ship something real.
Next cohort: July 6 to 17, 2026. Austin, TX. small cohorts.
Can’t take two full weeks off?
A 4-week part-time format. Fridays plus weekends. Same curriculum. Same three shipped products. Same public Demo Day, built around your work schedule.
Day by day. Same three builds.
Set up the tools. Leave with a PRD.
How LLMs actually work. The real differences between Claude, GPT, Gemini, and DeepSeek. Talking to Claude like an expert, not a search engine.
Warp Terminal, first GitHub push. First look at Claude Code in the terminal.
Write your capstone PRD with Claude as co-author. Scope it small enough to finish, real enough to use. Pushed to GitHub by 3pm.
Build a live site. Ship Product 1 Sunday.
Build a real landing page in plain English. Connect GitHub to Vercel. Push to main, auto-deploy in 30 seconds. Live URL by end of day.
Databases without jargon. Set up a Supabase table, connect to your Vercel site, query it with Claude, ship a simple dashboard.
Build your first Claude Skill. Learn how AI connects to your tools. 4-min Product 1 demo to your cohort on real data.
Connect Claude to the real world. Ship Product 2.
Pipelines vs. agents. Connect Claude to Gmail, Slack, HubSpot, and 200+ tools via Composio — no integration code. Build automations on your real data.
Playwright for sites with no API: navigate, extract, schedule. Voice stack: Whisper transcribes, Claude reasons, ElevenLabs speaks.
Claude Code as a full project collaborator in Warp. Ship a working AI agent that connects to an external tool. Deployed on Render, running on your own data.
Capstone sprint. Demo Day Sunday.
Evals, document Q&A, and data/access safety. Capstone build kicks off — no new scope after 3pm.
No lectures. No new scope. Build. Capstone deployed by 2:30pm. Dry-run demos with peer feedback. Backup screen recording for every demo.
Austin tech community in the room. 5 minutes each plus Q&A. Live demos on real data, hard questions, real feedback from builders and investors.
Get the 4-week curriculum
Download the full breakdown of our part-time format and get updates as we open this new program.
What it feels like to learn this.
“Code seemed like a different world I could never connect to mine. By day two, working in the command line, I felt infinitely more comfortable — I got everything I knew I needed, very quickly.”
Ted Kilgore · Cohort 1 feedback session
“The most valuable part was building the base — then taking that base to build the agent.”
Cameron Dowd · Cohort 1 feedback session