Personalized Learning Path
A personalized learning path is a sequence built for one person's goal, level, and time, not a catalog you browse. The U.S. Department of Education's 2016 National Education Technology Plan defines personalized learning as instruction in which the pace and the approach are optimized for each learner. Objectives, approach, and content, including its sequence, may all vary with that learner's needs, and the activities should be relevant to them and are often self-initiated (Future Ready Learning). The sequence you write includes what this person can skip.
Richard Culatta, then director of the Office of Educational Technology, separated the neighboring terms in EDUCAUSE Review. Adaptive learning means technology assigns the next resource from a learner's responses to questions, tasks, and experiences. Individualized learning keeps the same experience and only changes the speed, which is how Culatta describes most massive open online courses. A personalized learning path changes which modules are on the list.
Weekly review, a short resource list, and adjusting the plan as you go are covered in how to design a personal learning roadmap. That page uses the phrase personal learning roadmap. This page is the sequence.
The outcome
Write what the person should be able to do, in a form someone else could check. "Learn Python" names a topic. "Write a script that reads last month's CSV and prints the three customers with the largest unpaid balance" names an outcome. The Department of Education definition puts objectives first because a sequence has nothing to follow until the objective is specific. A module that does not change whether this person can do that thing does not belong on the path.
What they already know
David Ausubel opened Educational Psychology: A Cognitive View (1968) with one principle: the most important factor in learning is what the learner already knows, so find that out and teach from it. Joseph Novak quotes that line as the reason to identify the concepts this learner already holds before adding new ones (Novak's account).
For one person, the list is short. Name the tools they already use, one skill they can perform today, and one task they tried and could not finish. That list is what you delete. Someone who already writes SQL joins does not need a module on what a table is. Someone who has never opened a terminal does not start at a deployment chapter. A label such as "intermediate" is weaker than the list, because two people with the same label are missing different pieces.
The order of modules
Robert Gagné's learning hierarchies treat an intellectual skill as dependent on simpler skills that should come first. You find those skills with a task analysis: start at the outcome and ask what the person must already be able to do in order to attempt it. Those answers become the earlier modules. The hierarchy is a basis for the sequence (Conditions of Learning).
For the unpaid-balance script the order is: read a CSV in Python, filter rows, sum by customer, sort, and print the top three. Authentication and a survey of data science are other outcomes. If you cannot name the later step that needs a module, cut it. This orders skills that depend on earlier skills. It is not evidence that one order beats every other order.
The weekly cap
Pace is part of the same Department of Education definition, so the hours belong in the path. Write a weekly cap in hours this person will actually spend, then stop adding modules when the sequence no longer fits. Four hours a week for six weeks can hold a handful of small modules. It cannot hold a survey of the field. If a module will not fit, drop it or move the outcome closer. A path that assumes ten hours from someone who has three describes a different person. Hours per week for a self-paced course is a way to split a cap into sessions.
What you are not handing to software
Culatta's adaptive learning picks the next resource from responses. A review of adaptive recommenders from 2015 to 2020 describes a further step: the learner names a target topic, and a model recommends a sequence along a knowledge graph of concepts (Raj and Renumol). Conati and Merten used eye tracking to model self-explanation during an exploratory math tutor, so the tutor could tailor support in the session (their evaluation). Writing the outcome, prior knowledge, order, and weekly cap does not require either system.
Culatta also marks learning styles as unsupported. One helpful video does not make someone a video learner. Match the activity to the skill: a script outcome needs the person to run the script.
Where Ailurn fits
Ailurn turns a prompt into a course you take in the same workspace. In planning chat you state the outcome, what you already know, the module order, and the weekly cap. That prompt is the personalized learning path. The course is generated from it.
You can paste a documentation URL into the same chat to ground the outline. On a paid plan you can attach a PDF. GitHub repository ingestion, which reads the repository rather than only a URL typed into the chat, is on Plus. Free is planning chat, a full outline, and one lesson from a prompt. File attachments start on a paid plan.
Ailurn does not infer the next module from clicks, eye tracking, or a knowledge graph. It does not watch a YouTube video, and pasting a link does not import a transcript. It does not issue an accredited certificate. When the goal is a credential an employer already names, use the program that issues that credential.
Start from the course builder
Open the AI course builder and put the four parts in one prompt: the outcome, what you already know, the order of modules, and the weekly cap. You get that personalized learning path as a course, and you take it in the same workspace.