Interactive training · Program details
Master the AI wave. Become Agentsable.
The Agentsable interactive training is a four week program for mastering autonomous AI system architecture and deploying your first production ready agent fleet. Everything you need to decide whether it fits is on this page, including tuition and what it does not do.
The outcome
By the end of week four you have a working agent fleet solving a real problem from your own work, deployed, monitored, with a written failure profile, and you can explain why every part of it is shaped the way it is.
That is a planned outcome for participants who complete the work, not a guarantee. What we control is the structure, the review, and the honesty of the feedback.
How it works
Day Zero project
Your own problem, chosen on day one, carried through every session.
Live labs
Weekly working sessions where you build and break things with others present.
Guided review
Structured critique of your decisions, plus reviewing two other participants' systems.
All of it runs on the six stage Agentsable cycle, theory, real work, immediate application, failure inspection, feedback, revision.
Four weeks
Week 1
Day Zero project and mechanism
You pick a real problem from your own work on day one and it stays with you for four weeks. Alongside it: how agent loops, tools, memory, and retrieval actually behave, taught at mechanism level.
Week 2
Architecture and decomposition
Breaking your problem into steps a system can attempt and you can supervise. Where checkpoints belong, what gets escalated, what must never be automated.
Week 3
Failure, evaluation, and guardrails
You deliberately break your own build, write the diagnosis, and add the cheapest checks that would have caught it. Guided review of two other participants' systems.
Week 4
Deploy the fleet and defend it
Your agent fleet goes to a production ready state with monitoring and a documented failure profile. You present it and defend the decisions, not the demo.
Evidence
This is an early cohort. We have delivery experience behind the method and small group runs behind the format, but we do not yet have participant artifacts or testimonials we are permitted to publish. Rather than borrow credibility, we are saying so.
What you can inspect today: the learning method, the worked exercise, and the origin and limitations page. Judge us on those.
Logistics & tuition
- Format
- Interactive training, on demand lectures, live Zoom Q&A, live RAC builds, and hands on browser labs
- Duration
- Four weeks
- Weekly commitment
- 8–10 hours
- Cohort size
- Capped at 100 participants
- Requirements
- Comfortable reading code and using a terminal; a real problem to work on
- Tuition
- To be confirmed before enrollment opens, no charge is taken until it is published here
- Next cohort dates
- To be announced. We will not display a date until it is fixed.
- Refunds
- Full refund up to 7 days before start; partial after, stated in writing beforehand
Fit, questions, enrollment
This is for you if
- You already build or ship things and want AI judgment inside that work.
- You have a real problem you can bring on day one.
- You would rather be told your design is wrong than be encouraged.
- You can protect 8–10 hours a week for four weeks.
This is not for you if
- You want a survey of AI tools or a certificate for your profile.
- You have never used a terminal.
- You cannot attend most live labs.
- You are looking for a guaranteed career outcome. We do not offer one.
Direct answers
Do I need to be able to code?
You need to be comfortable reading code and running things in a terminal. You do not need to be a professional engineer. If you have never opened a terminal, this cohort will be painful, start elsewhere and come back.
How much time does it take each week?
Plan for 8–10 hours: roughly 3 hours of live labs, the rest on your own build. People who spend less than 6 hours generally do not finish the Day Zero project, and finishing it is the point.
Do I have to attend live?
Labs are the part that does not work asynchronously, the review and the failure diagnosis happen there. Sessions are recorded, but attending at least three of four is the honest requirement.
Can I use my employer's real data?
Only if you are authorised to. Most participants use a realistic synthetic version of their problem, which works fine for everything the programme teaches.
What exactly do I leave with?
A deployed agent fleet solving your own Day Zero problem, its documented failure profile, the evaluation checks around it, and the reasoning behind each architectural decision.
What is the refund policy?
Full refund up to seven days before the cohort starts, no questions. After week one begins, a partial refund is available at our discretion, we will tell you the amount before you commit, in writing.
Is there a completion certificate?
There is a completion record listing what you built. We do not pretend it functions as a credential, because it does not.
Join the waitlist
Enrollment is not open yet, so there is nothing to pay and no seat counter to watch. Join the waitlist and you will get the dates and tuition when they are fixed, before anyone else is asked to decide.