From AI user to AI professional
An AI professional can define a problem, choose the simplest capable system, ground it in evidence, test failures, protect people and data, ship it, and explain the value. Tool familiarity alone is not the finish line.
7 complete free courses · 126 public lessons · no income or job guarantee
The sequence
Learn in four stages
Do not collect courses randomly. Finish each stage with evidence another person can review.
Control the model
Ready nowLearn context, source boundaries, structured briefs, examples, task decomposition, and review. The goal is repeatable work—not clever prompt tricks.
Proof before progress
A tested prompt package and a reusable assistant for one real task.
Build reliable AI systems
Ready nowMove beyond chat into APIs, JSON, tool contracts, data, multimodal inputs, retrieval, agents, evals, security, logs, and deployment.
Proof before progress
A deployed workflow with a README, test set, eval results, threat model, cost log, and demo.
Choose one applied specialty
Ready nowDepth wins. Choose the job, creator, video, or content lane that fits your existing strengths, then complete its real capstone before adding another lane.
Proof before progress
Two portfolio artifacts for the same audience or problem, each with sources and a QA checklist.
Run an ethical earning experiment
Prepared flagshipPackage one useful outcome, validate demand, use written scope and payment milestones, deliver with human QA, and ask for consented proof or repeat work.
Proof before progress
An offer, portfolio sample, discovery notes, proposal, scope, delivery checklist, and 30-day activity review.
Professional core
Ten competencies—not ten tool logos
The current builder core reflects the disciplines in official API, retrieval, agent, evaluation, and safety guidance.
Problem framing and acceptance criteria
Source verification, research, and citation discipline
Prompt, context, and reusable instruction design
JSON, HTTP, APIs, structured outputs, and tool contracts
Spreadsheet, document, image, and audio workflows
Retrieval, grounded answers, permissions, and freshness
Agents, state, retries, idempotency, and human approval
Evals, red-team cases, logging, cost, and reliability
Privacy, security, consent, copyright, and disclosure
Portfolio proof, discovery, scope, delivery, and review
Earning paths
Choose how the skill creates value
Income is an outcome of useful work, proof, fit, timing, and execution. The curriculum teaches controllable steps—not a guaranteed number.
Employment
Proof to build
A role-specific evidence pack and a small workflow relevant to the target job.
First honest test
Apply selectively, practise interviews, and measure application quality—not a promised offer.
AI-assisted service
Proof to build
Two honest samples, a narrow offer, written scope, and a human-QA checklist.
First honest test
Validate the problem, propose a small paid pilot, and record delivery evidence.
Builder or automation work
Proof to build
A deployed workflow with tests, logs, permissions, cost notes, and a clear handoff.
First honest test
Solve one bounded workflow before proposing agents or a large automation stack.
Creator or digital product
Proof to build
A content or product sample tied to a real audience question and transparent claims.
First honest test
Publish a small experiment, review first-party evidence, and validate before expanding.
A realistic sprint
Your first 90 days
Adjust the pace for work, study, power, and connectivity. Completion quality matters more than a calendar promise.
Days 1–30
Foundation
Finish AI Fundamentals and Advanced Prompt Engineering. Save tested outputs, not screenshots of completion.
Days 31–60
Build
Complete Applied AI Builder and one specialty capstone. Publish two honest portfolio artifacts with limitations.
Days 61–90
Validate
Choose one earning lane. Run interviews, applications, audience tests, or a small paid-pilot proposal and review the evidence weekly.
Source desk
Built around current professional practice
Product interfaces change. The core course teaches transferable concepts and points current implementation details back to official documentation.
What this roadmap does not promise
It does not promise a salary, client, audience, certification equivalence, or a fixed completion time. It gives you sequenced practice, review criteria, and portfolio proof. Employers, clients, platforms, and markets make their own decisions.
Build an earning system