Bring an AI project. Leave with a decision you can defend.
The method is in my free book. Read it first.
The course is four weeks to deep dive into your project, covering one part of the loop each week, so you walk away with a framed out project.
🚨 Winter cohort: Jan 4, 2027 🚨
Get 20% off with code: Earlybird
Two audiences:
🏗️ You are about to build. The course helps you frame it, write a first-slice spec, hand it to a coding agent, read what it got wrong.
🩺 Or your project is stuck. We will combine framing with diagnostics to figure out how to improve your project.
What you get:
🧑🏫 Four live sessions, 75 minutes each, your project on screen
📄 Four one-page artifacts on your project: Framing Brief, GOATS One-Pager, Diagnostic Readout, Decision Memo
🛠️ The hands-on exercises and worked examples that aren't in the book
🤖 The frame-use-case and diagnose-use-case skills for coding agents
🎥 Lifetime access to recordings
This course is for you if:
• You have an AI idea and want to make sure it's done right
• Your AI project started and it's not looking good
• You have to communicate decisions around your project that might be changing course or stopping
More at aiframer.dev.
What you’ll learn
One part of the GOATS loop each week, on your own project.
A framework to start with: a 5-step GOATS framework (Goal, Operating Assumptions, Alternatives, Trade-offs, Signals)
Catch the underspecified setup that sinks projects: the wrong metric, the wrong atomic unit, the easy assumptions.
Tell a model problem (keep tuning) from a framing problem (rethink the setup) instead of guessing.
Spot the warning signs in ML, GenAI, and agent systems before they get expensive.
Learn how to set success and stop criteria for your AI projects
Reframe a stuck project into one that can actually work.
Question requirements before you build them. Push back with evidence: "I know you want a chatbot, but here's why search is better
Who this course is for
You've shipped AI and walked away frustrated. You know it could have gone better, and you don't want to run the next one on instinct.
You shipped something, the demo worked, momentum stalled, and you can't tell whether it's the model or the framing.
You came up through engineering, product, or data, and you have to tell someone whether to keep going, change course, or stop.
Prerequisites
You understand what AI is at a conceptual level.
We focus on framing and strategy, not explaining what models or training means. Basic vocabulary lets us go deeper faster.
You bring a real project
The frameworks click when you can map them to something you are building. Without that, they stay abstract and it's harder to learn.
You can run a coding agent for one hour
Coding agents are great for getting feedback and questioning assumptions on your projects. We will learn the best way to use them.
What's included
Live sessions
Learn directly from Rajiv Shah in a real-time, interactive format.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Course syllabus
4 live sessions • 68 lessons • 6 projects
Live sessions
2 hrs / week
One 75 minute session a week, your projects on screen. I work through some examples, then you do yours. Recorded.
Projects
1 hr / week
One artifact a week on your project.
Recorded Lessons
1-2 hrs / week
The book chapter for the week. Recorded video lessons from earlier cohorts are there if you want more.
Week 1
Jan
9
Session 1
Lesson 1: Thinking from First Principles for AI
Bonus: AI Alternatives for Gen AI and Traditional ML
Week 2
Jan
16
Session 2
Lesson 2: The GOATS Loop, A Framework for AI Problem Framing
Free resources
Why this matters
"The key question for enterprise AI is: what problem are you actually solving?" -- Alexandru Vesa, MLOps Engineer
"Tools go in and out of style quickly, but fundamental approaches to problem solving should last a bit longer." -- Chip Huyen, Stanford ML Systems Design
"The future of work is all of us becoming managers of AI." -- Richard Socher, former Chief Scientist at Salesforce
Course Map
The Outer Loop for Reframing Your AI Projects
What You will Learn
The Strategy & Mindset
- The 5-Step frameworks: GOATS is structured engineering lifecycle for deconstructing and reframing complex AI problems before writing code.
- Thinking Methods: Inversion, De-escalation, and the Feynman Test. Break System 1 pattern-matching and catch bad framing before it becomes expensive.
The Diagnostics
- Diagnostic Tests: Figure out whether your agent problem is model capability, orchestration, task framing, or an alternative approach.
- Failure Funnel: Identify where errors are happening.
- Pivot Signals: Tests for recognizing when an approach has hit a dead end (and exactly what to do about it).
The Toolkit
- Canvases & Checklists: Course-exclusive worksheets for running strategy sessions, including the AI Problem Framing Questionnaire.
The Reference Library
- 200+ Case Studies: A searchable database of real-world pivots (Netflix, Uber, Stripe) mapped to the 18 architectures in the course.
- Manual: A 250 page book that puts all the lessons into an easily readable form
Participation & Code of Conduct
Participation Expectations
This is a practitioner-focused cohort built on trust and discussion.
Participants are expected to engage respectfully and professionally.
Disruptive behavior, bad-faith participation, or misuse of course materials may result in removal.
Course content, discussions, slides, and materials are for personal educational use only (I handle group corporate training differently)
If you have any questions, reach out. I spend a lot of time preparing and want to make sure we have a conducive environment for learning.
Frequently asked questions
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort$450
USD