Enrolling now · Start any day

Learn the AI stack. Build for a real US business. Leave with their name on your résumé.

A self-paced online course for $300: most people finish in about 12 weeks at 2 hours a day, with no deadline. You learn to build and run AI agents with small LLMs, Hermes, OpenClaw, Jev and agent harnesses, then spend 1 to 3 months as an intern building a real system for a US business. Many owners want the person who built it to keep building: as a Forward Deployed Engineer on staff, on contract, or on a paid monthly retainer. It is a chance, not a promise. No coding background needed.

Takes 3 minutes. Nothing to pay to apply. $300 only once you are accepted. Start the same week.Continues with your Google account. No Google? Apply with your email below.Not sure yet? No application, no payment: module 1 in full and a preview lesson from every other module.How the programme works, step by step · See pricing and plans
$300all inHiring optionat the end of every internship
Your mentor

The classifier keeps putting refund requests under general questions. What should I check first?

Pull the mislabelled emails into a test set and compare them with the examples in your prompt.

Example internship briefWeek 3 of 6
Bookkeeping firm, Austin TX, 12 peopleBuild an agent that sorts our client inbox and drafts replies for review.
Small LLM (local)Hermes agentOpenClaw + GmailJev decisions
  • Week 1: scoped the problem with the business owner
  • Week 2: inbox classifier running on a local model
  • Week 3: draft replies, owner check-in Thursday
AIMentor: AI mentorTrainee: [Your name]
Offer: Forward Deployed EngineerOne possible outcome at week 6. Every intern leaves with a reference letter.
The problem

You already know AI is where the jobs are. Nobody tells you how to get your first one — or where.

The Catch-22

Every job post says “AI experience required”. But no one will give you experience until you have experience.

Tutorial Hell

You have watched the tutorials. You still could not ship a working agent if someone asked you to tomorrow.

The Certificate Trap

Certificates get ignored. “I built this for a company in Texas, and they kept me on” gets a reply.

How it works

Learn it. Build it. Then do it for a real business.

  1. About 12 weeks at 2 hours a day

    Learn the stack, at your pace

    Online, self-paced, no deadline. You run a small LLM on your own laptop in the first lesson and add Hermes, OpenClaw, Jev and a harness module by module. Go faster if you have the hours, slower if life gets in the way.

  2. Your capstone

    Build your own agent

    You pick a task from your own life or work and build an agent that does it end to end. It has to run before you move on — this is what earns you a placement.

  3. 1 to 3 months

    Intern with a US business

    We assign you a US business with a real problem. You scope it with the owner, build it with help from an AI mentor, and hand over a working system. Most owners then ask the obvious question: can you build the next one for us?

See the outline, lesson by lesson

You also learn to give an agent a memory a business can trust, in Agent memory, and to set up and grow a business's own apps on its own server, in Own your stack: the Business Suite.

What you will be able to do

Not “introduction to AI”. The stack people are actually building with right now.

Each module ends with something that runs. If you cannot demo it, you have not finished the module.

  1. Small LLMs

    Run a capable model on your own laptop, offline

    Pick the right open model for a task, run it locally, and know when a small model is enough and when it is not.

  2. Hermes

    Build an agent that plans, remembers and uses tools

    Set up Hermes agents, give them memory and skills, and run more than one at a time on a small server.

  3. OpenClaw

    Give your agent hands: email, browser, files, chat

    Connect an agent to the apps a business already uses so it can act, not just answer.

  4. Jev

    Make an agent decide well, not just chat

    Use Jev to turn messy inputs into consistent, explainable decisions you can log and review.

  5. Harnesses

    Wire it into a loop that finishes real tasks

    Planner and sub-agents, retries, checks and hand-offs, so a job gets done even when a step fails.

  6. Ship it

    Deploy for a business and hand it over cleanly

    Put it on a server, write the one-page handover, walk the owner through it on a call, and keep them informed with stages and build items.

The library

Build with the real AI stack. No sandbox fluff.

These numbers come straight from the library, so they go up as we add workflows, connectors and skills.

Business types
53
Kinds of small business with workflow designs written for them, in 10 groups.
Workflows
664
Designs for one job an agent can do: what starts it, the steps, the tools and the human check.
Connectors
192
Systems an agent can work with, each with how it connects and what it may do there.
Official MCP servers
59
Connectors whose maker publishes an official MCP server, the standard way an AI agent plugs into a system.
Skills
82
Packaged instructions so an agent does one job the same way every time.
Tools
10
The agent and automation tools interns build with: the connectors listed under AI models and agents, and under automation platforms.

Every connector and skill in the library is approved for client work.

The internship

A real US business. A real problem. Your name on the solution.

The businesses sign up on this site. They bring a problem they actually want solved. You get matched by interest, scope it on a call with the owner, and build it over 1 to 3 months, depending on the project, with an AI mentor you can ask at any hour and a check-in with the owner every week. Businesses run several interns in parallel and are paid for completing internships properly, with an honest written review, which is why they review your work closely.

The internship is unpaid, and a job at the end is a chance, not a promise. We say both up front. What every trainee takes home is a working project, a reference letter from a US business owner, and a portfolio piece that answers the “do you have experience?” question for good. Some also take home an offer.

  1. Get matched

    We pair you with a business whose problem fits what you enjoyed building most in the course.

  2. Scope it with the owner

    One call. You write the brief. You learn how to ask a business what it actually needs — a skill most engineers never get.

  3. Build, review, repeat

    Weekly check-ins with the business, and an AI mentor on your internship page whenever you are stuck. You ship something that works every week, not a big reveal at the end.

  4. Hand over. Get your letter. Maybe get the job.

    You walk the owner through the finished agent. They sign your reference letter. And because they have watched you build for months, the question “can you keep going?” is a natural one to ask. Many owners do; it is a chance, not a promise. They can also refer you to other businesses in one click, and the referral carries proof of your work: your brief, your weekly log, your approved capstone and their reference letter. You are told every time one is sent, and you can withdraw it whenever you like.

Four things can happen, and more than one of them can.

  • Hired as staffThe business hires you as staff.
  • Kept on contractIt keeps you on contract for the next pieces of work.
  • Reference letterIt signs your reference letter, which every intern leaves with.
  • ReferralIt can refer you to other businesses in one click, with proof of your work attached.

A job, a contract or a referral is a chance, not a promise.

Referral[Your name]Hired
The briefBuild an agent that sorts our client inbox and drafts replies for review.
Work logWeek 1: scoped the problem with the business ownerWeek 2: inbox classifier running on a local model
Reference letterSigned by the business owner
Example proof-of-work page. A referral opens it without an account.
After the internship

Then it gets paid. Monthly retainers.

When your internship is completed, the business can keep you on a monthly retainer: 30 hours of your work each month, on the system you built and the next ones it needs. The internship itself stays unpaid; the retainer is paid work.

  • You keep 80% of every retainer

    The business pays $750 a month for each retainer to us, the contracting party, and we pay you 80% of it.

  • Up to 4 at once

    You can hold up to 4 retainers at the same time, each with a business that has already watched you work.

  • Our gateway, and a bench behind you

    You work through our gateway, as in your internship: the owner approves what leaves their account, and every request is logged. If something urgent comes up when you cannot take it, the bench answers.

  • We handle the contract and the billing

    We sign the retainer with the business, bill it every month and pay you your share. You never have to chase an invoice.

An example at today's numbers, not a promiseAt $750 a retainer, with a share of 80% for you. Each retainer is 30 hours of work a month.
What you would keep each month. An example at today's numbers, not a promise.1 retainer: $600 a month for 30 hours; 2 retainers: $1,200 a month for 60 hours; 3 retainers: $1,800 a month for 90 hours; 4 retainers: $2,400 a month for 120 hours1 retainer30 hours a month$600 a month2 retainers60 hours a month$1,200 a month3 retainers90 hours a month$1,800 a month4 retainers120 hours a month$2,400 a month

A retainer is the business's choice, offered when the internship is completed. Like a job or a referral, it is a chance, not a promise.

The job at the end

Businesses do not need another AI course graduate. They need someone who can make AI work in their business.

That person is a Forward Deployed Engineer: someone who takes a small business's messy, real processes and builds agents that run them. Most small US businesses cannot hire one — the people who can do this are expensive and rare. After one to three months of watching you do exactly that job for them, you are neither.

What the role looks like

Remote. You own the agents you built and the next ones the business needs: new integrations, fixes when a model or an app changes, small automations that save the owner hours a week.

Why the internship is the interview

A CV gets you ten minutes. The internship gets you months of the owner seeing your work every week. By the end there is nothing left to assess — only a second project waiting.

How interns become staff

The first agent is never the last. Once an owner has one automation working, they see the next five. The person who built the first one is the obvious person to build them — that is how most of our interns get absorbed. It is not guaranteed: a business may lack the budget or the fit. Every internship still ends with a reference letter and a shipped project.

Who joins

Three kinds of people. One thing in common: they want proof, not promises.

Students

Problem
Final year or just graduated, and every classmate has the same degree.
Edge
You will have the degree and a US business that vouches for you.
You leave with: a US reference, and a real shot at your first job before you graduate.

Homemakers returning to work

Problem
A career gap that every interview asks you to explain.
Edge
The course is online and self-paced, and the internship is remote on a schedule you agree with the business. The gap ends with a current project, not an explanation.
You leave with: recent, relevant work on your CV.

Working professionals

Problem
A full-time job, and no time for classes or a fixed weekly call.
Edge
Self-paced in your evenings and on weekends; the US owner reads your weekly check-in and approves your work on their dashboard. Come back to your team as the person who can actually build agents, or don’t come back, if the business makes you an offer.
You leave with: a shipped agent and a story to tell your manager.
Questions

Things people ask before they apply

I have never written code. Can I do this?

Yes. The first two weeks teach exactly the amount of Python you need, and the tools you use the rest of the course mean you write far less code than you would expect. You need a laptop and about 2 hours a day if you want to finish in 12 weeks — or fewer hours over a longer stretch. There is no deadline.

Is the internship really unpaid?

Yes. The business pays nothing and you are paid nothing. That is what makes it possible for a small US business to hand a real problem to someone with no track record. You get the work, the reference and the portfolio piece.

Is it online or in person?

Everything is online. There is no center and no classroom, so you can join from anywhere with a laptop and a steady connection. The course is self-paced: lessons, exercises and an AI mentor you can ask at any hour. The internship is remote, with calls scheduled around the US time difference.

Can I do it alongside a job or family?

That is who it is built for. The course is self-paced with no time limit, so you set the hours. The internship is remote and runs 1 to 3 months on a schedule you agree with the business; expect it to need more of your week than the course did.

Can I try it first?

Yes. Start free: module 1 in full and a preview lesson from every other module, with the AI mentor for a few questions a day. No application and no payment. When you are ready, apply from the same account and your progress carries over. A free account gives no right to an internship placement. Start free with module 1

What does it cost?

$300 for the whole program — the course, the AI mentor and one internship placement. Applying is free. You pay only after you are accepted, and nothing else is required. The fee covers one placement. Want more depth? Add a second or third business for $50 each, at the time you pay or later. Full refund within 14 days of paying if you have not completed the second module, or if we cannot offer you an internship within 90 days of your capstone approval. No refund once an internship starts. Each extra business you add is refunded on its own, in full, if you cancel it before we match you for it or we cannot offer it within 90 days of it becoming due. The refund policy has the details.

Who is the mentor?

An AI mentor, not a person. Once you have scoped the project with the owner, it lives on your internship page and answers your questions at any hour. Your capstone is reviewed by AI against a fixed rubric, and we can step in and change the result. The business owner is the person you speak to every week.

Will I get a job at the end?

There is a real chance, not a guarantee. Many owners want the person who built their AI system to keep building for them: they can hire you as staff, take you on contract, or keep you on a monthly retainer through us. We cannot promise a job and we will not. What every trainee leaves with is a working project, a reference letter and a portfolio piece.

What happens after the internship?

Four things can happen, and more than one of them can. The business hires you as staff. It keeps you on contract for the next pieces of work. It signs your reference letter, which every intern leaves with. And it can refer you to other businesses in one click, with proof of your work attached: your brief, your weekly check-in log, your approved capstone and their reference letter, on a page the other business can open without an account. You are emailed every time a referral is sent, you see every one on your internship page, and you can withdraw any of them at any time. A job, a contract or a referral is a chance, not a promise.

How do retainers work?

When your internship is completed, the business can offer to keep you on a monthly retainer: 30 hours of your work each month, looking after the system you built and building the next pieces. The business signs the retainer with us and pays us; we are the contracting party, so we handle the contract and the billing. You work through our gateway, as in your internship, and a bench of other trained interns answers urgent requests when you cannot. You can hold up to 4 retainers at once. It is the business's choice, so a retainer is a chance, not a promise.

What does the intern keep from a retainer?

80% of it. The business pays $750 a month for each retainer, and we pay you 80% of every month that is billed. The internship itself stays unpaid. The worked example on this page multiplies the share by each retainer you could hold, at today's numbers; it is an example, not a promise of income.

How do you choose the businesses?

Every business applies on this site and speaks with us before they are listed. We only accept problems that are real, scoped to 1 to 3 months, and safe to give to a trainee working with an AI mentor.

Run a US business?

Get an AI agent built for a problem you actually have, by a trained intern with an AI mentor, at no cost.

Post a project

Twelve weeks of learning. A few months of real work. Then an owner asking you to build the next one.

Self-paced and online, so you can start any day. Placements go to trainees whose capstone agent works, in the order they finish.

$300, paid only after you are acceptedWe reply within 2 working days with a short callFully online — join from wherever you are

The fee covers one placement. Want more depth? Add a second or third business for $50 each, at the time you pay or later.

Full refund within 14 days of paying if you have not completed the second module, or if we cannot offer you an internship within 90 days of your capstone approval. No refund once an internship starts. Each extra business you add is refunded on its own, in full, if you cancel it before we match you for it or we cannot offer it within 90 days of it becoming due. Refund policy · See pricing and plans

Start free with module 1 No application, no payment: module 1 in full and a preview lesson from every other module.

Your background (optional)

Helps us match you with a business that fits. Not shown anywhere on its own.