// Instructional Design · Case Study
Teaching
Myself
to Teach
Four very different rooms — K-12, truck drivers, the DOD, NCAA athletes — welded into one platform that works in person and at midnight. The LAZRdojo LMS.
The Story
I did not set out to become an instructional designer. I set out to make things.
What happened along the way is the thing that happens to a lot of self-taught makers, though most people don't stop to notice it: in order to build what I wanted to build, I had to learn far more than anyone had asked me to learn. Nobody assigned it. There was no syllabus. Every project opened a door onto three more skills I didn't have, and I went and got them, because the project wouldn't exist otherwise.
Somewhere in that process I started paying attention to how I was learning rather than just what. The pattern was consistent, and it did not look like school. It looked like this: want something badly enough to start it, get a rough version working, show it to someone who knew more than me, hear what was weak, go make it better. Repeat until it was good. Nobody quizzed me. Nobody gave me a deadline. The project was the assignment, the critique was the grade, and my own investment in the thing was the entire motivational engine.
Then I taught. Not one population — four, and they could not have been more different: K-12 students, truck drivers, personnel at the Department of Defense, and NCAA athletes.
Nothing transfers between those rooms on the surface. A twelve-year-old, a career driver, a defense analyst and a Division I athlete share no schedule, no vocabulary, no motivation, and no reason to be there. What became obvious across all four was that the things which actually worked were the same every time, and they were never the things a course catalog would have predicted. Show the work being done rather than described. Let them try it badly first. Say one specific thing about what they made. Let them come back to it later.
What I wanted to build was something that welded all of it into one place — a platform that works the same whether the person is standing in front of me or working alone at midnight. And I wanted it to behave like a sketchbook: something an artist opens to see their own progress, keeps as a reference, and goes on learning from long after the page was filled. Not a transcript. Not a gradebook. A record of becoming better at something, owned by the person who made it.
So I went and studied the field: how learning is actually structured, what the research says about feedback, motivation, assessment, and transfer, and where the formal discipline of instructional design agreed or disagreed with what I'd arrived at in those four rooms.
What I found was that most of what I'd arrived at by building and by teaching had names. Studio critique. Cognitive apprenticeship. Mastery learning. Authentic assessment. Constructionism. I had reinvented a good deal of the field the long way around, by doing it. What the literature gave me was precision — the ability to see which parts of my instinct were load-bearing, which were accidental, and which were quietly working against me.
LAZRdojo is what came out of that. It's a creative-technology program for young makers, and it runs on a learning platform I designed and built end to end. The platform is not a container for the pedagogy.
expressed in a schema.
This page is the argument for that claim.
The Instructional Model
Three commitments sit underneath everything. Each is a position I could defend in a room full of instructional designers, and each one shows up in the code.
01 — Framework
Cognitive apprenticeship, not instruction
The dominant model here is the one Collins, Brown and Newman described: make expert thinking visible, then hand it over gradually. A mentor models the work, coaches while the creator attempts it, scaffolds heavily at first, and withdraws that scaffolding as competence grows. The creator articulates what they're doing and reflects on how it went.
This is why the substantive feedback in LAZRdojo does not live in the software. It happens in conversation between a creator and a mentor, in the moment, over the work. I made that choice deliberately and I would make it again. Automated feedback scales, but it can only respond to what it was built to anticipate. A mentor looking at a half-finished build can see the thing the creator doesn't know they don't know — and that is the entire value of an apprenticeship.
The platform's job is to make the apprenticeship possible: hold the structure, track the progression, and get out of the way of the conversation.
02 — Assessment
Studio critique as the assessment engine
Every deliverable goes through the same two-beat rhythm that governs any working studio.
→ First, get the thing done.
Not perfect. Done. A finished rough thing is a real object you can talk about; an unfinished perfect thing is nothing.
→ Then, can it be better — and how?
Small, specific, iterative critique. Not a grade. Not a rubric score. A conversation about one or two things that would make this particular piece stronger, and then the creator goes and does it.
This is the atelier model, and it's how every serious creative discipline has trained practitioners for centuries. It's also, in the language of assessment research, high-quality formative feedback: specific, timely, focused on the work rather than the person, and actionable while the work is still in progress. The reason it isn't encoded as a rubric is that a rubric fixes the criteria in advance, and in creative work the interesting criteria emerge from the piece itself.
03 — Pacing
Mastery learning — time varies, standard doesn't
Most things in LAZRdojo have no deadline, and everything can be revisited.
That's a direct implementation of what Bloom called mastery learning, built on Carroll's insight that aptitude is largely a measure of how long a learner needs, not whether they can get there. Conventional courses hold time constant and let achievement vary — everyone gets six weeks and comes out at whatever level they come out at. Mastery learning inverts it: hold the standard constant and let time vary.
The motivational consequence matters as much as the pedagogical one. A system without deadlines and with unlimited revision has a mastery goal orientation rather than a performance one. Creators aren't optimizing for a score under time pressure; they're optimizing for the thing being good. That produces persistence in the face of difficulty rather than avoidance of it — the single behavior that determines whether a young maker keeps going.
What replaces the deadline is the creator's investment in their own project. That's the bet the whole program makes: a kid who genuinely cares about the thing they're building will out-work a kid who is being marked.
Built Into the System
Every task in every path carries one of four modalities. Point values scale with cognitive demand — which means a creator optimizing for XP is, structurally, optimizing toward creation.
Bloom's revised taxonomy compressed into a database field. Most platforms weight progress by seat time or quiz score, and consequently teach compliance without meaning to. Weighting by modality means the system pays out most for the work that cannot be faked.
Deliverable-based assessment
There are no quizzes anywhere in LAZRdojo. Every task terminates in an artifact: a screen recording, a playable build, a render, a written projection, a published portfolio. This is authentic assessment in Wiggins' sense — the evidence of learning is the same category of object a working practitioner produces. For creative-technical skill it's the only honest measure; a multiple-choice item can establish that someone recognizes the vocabulary of a thing they cannot do. It also makes the platform structurally a portfolio system with a progression layer, which is a more useful thing for a young creator to have than a transcript.
Backward-designed task authoring
Each task is authored deliverable-first: decide what the creator will produce, then write the description that gets them there. Backward design as a compositional discipline rather than a planning document — the schema won't let you write a task that doesn't terminate in evidence.
Editable path templates — differentiation with a real mechanism
Paths are authored as structured JSON: a path holds modules, modules hold tasks, and every task carries its modality, points, description, deliverable, and linked resources. New paths are built by forking existing ones — a mentor takes a proven template and edits it through a browser editor, no code or file transfer involved, into something specific to a creator or a group.
Two things make this significant. Customizing a path costs an instructor minutes, so differentiation is a practice rather than an aspiration. And the template library encodes accumulated pedagogical judgment: a new instructor doesn't start from a blank page, they start from a path that already works. The good decisions propagate.
Human placement and adaptive sequencing
There is no placement test. If a mentor talks to a creator and concludes the fundamentals are already there, tasks get skipped and the path gets altered. This is Vygotskian in the plainest sense — the target is the edge of what this specific creator can do with support, and the only reliable instrument for finding that edge is a person who has watched them work. Algorithmic adaptive sequencing infers the zone from click data. A mentor observes it directly.
Content decoupled from pacing
The same content objects run as a dated six-week cohort camp and as an open-ended self-paced membership, with no change to the underlying data. Pacing is a property of enrollment, not of curriculum. This is what the field spent two decades calling reusable learning objects and largely failed to implement, because most systems fuse the curriculum to its delivery schedule and then can't reuse either.
Progression, badges, and the motivation architecture
Read through Self-Determination Theory, the system supplies all three conditions Deci and Ryan identify for durable motivation. Competence: visible progression, accumulating XP, badges that fire on genuine milestones. Autonomy: creators write their own goals, mentors adapt paths to the individual, and the absence of deadlines returns control over pace to the learner. Relatedness: the mentor relationship, and cohort structure in camp programs.
On the perennial objection that extrinsic rewards undermine intrinsic motivation — the research distinction is between controlling rewards and informational ones. Badges awarded for attendance or compliance are controlling. Badges awarded for demonstrated mastery are informational: they tell a creator something true about their own growing capability. Every badge in LAZRdojo fires on a completed body of work.
Creator-authored goals
Each module opens with the creator writing three goals in their own words and closes with them reviewing honestly against those goals. Scaffolded prompts are available for younger creators, because a blank box is not autonomy — it's abandonment. This is metacognition built into the structure: goal-setting and self-evaluation are the two components of self-regulated learning most reliably associated with achievement, and the two most reliably left out of learning platforms.
The word "creator"
Participants are creators. Not students, not campers. That's an identity intervention, not branding — communities of practice research is clear that people learn a domain partly by taking on the identity of someone who works in it. Student describes a person receiving. Creator describes a person producing. What makes the word true rather than decorative is that the assessment model agrees with it. If I called them creators and then handed them quizzes, the vocabulary would be a lie, and young people detect that immediately.
Access designed around who the learners actually are
Authentication is a username and a four-digit PIN. No email address, no password reset, no verification loop, and no personal data collected from a minor beyond a first name. The pedagogical argument is stronger than the privacy one, though both hold: in a two-hour session, login friction is instructional time. A nine-year-old who cannot get into the system independently has been taught, before anything else, that this environment is not really for them.
Feature to Principle
| Feature | Principle | Effect |
|---|---|---|
| Modality field + weighted points | Bloom's revised taxonomy | Reward scales with cognitive demand |
| Deliverable on every task; no quizzes | Authentic assessment | Evidence is practitioner-grade work |
| Deliverable-first authoring | Backward design | No task without evidence |
| Mentor critique in conversation | Cognitive apprenticeship | Coaching responds to the unanticipated |
| "Get it done, then make it better" | Studio / atelier critique | Iteration as the unit of improvement |
| No deadlines; everything revisitable | Mastery learning | Time varies, standard holds |
| Open revision, no time pressure | Mastery goal orientation | Persistence over avoidance |
| Forkable templates, browser editor | Differentiated instruction | Customization costs minutes |
| Mentor-adjusted sequencing | Zone of proximal development | Challenge calibrated by observation |
| Creator goals + module review | Self-regulated learning | Learners direct and evaluate themselves |
| XP, badges, visible progression | SDT — competence | Growth made legible without control |
| Own pace, own goals, adapted paths | SDT — autonomy | Ownership of trajectory |
| Mentor relationship, cohorts | SDT — relatedness | Belonging in a practice |
| Making a real, shippable artifact | Constructionism | Learning through public making |
| Content decoupled from pacing | Reusable learning objects | One curriculum, many delivery models |
| "Creator" as the term of address | Communities of practice | Identity as a learning mechanism |
| Username + PIN, no email | Access as instructional design | Independence from minute one |
What I'd Build Next
Being honest about the edges is part of the discipline.
The gap I take seriously is spaced retrieval. A skill taught in week two and never revisited decays, and creative production programs almost universally skip distributed practice. Unlimited revisitation partly covers this, but it's opt-in — the creator has to choose to return. A deliberate callback structure, where later modules require earlier skills rather than merely permitting them, would close it.
Beyond that: capturing critique history against a deliverable so a creator can see their own trajectory of improvement, and lightweight peer critique, which would strengthen the relatedness leg for self-paced creators working alone.
What This Represents
I built this alone: the data model, the authentication, the instructor tooling, the content architecture, every learning path, and the design system it all runs on.
But the engineering is the smaller claim. The larger one is that every structural decision in it is an instructional decision I can name, defend, and trace to either evidence or deliberate experience — and that the ones I couldn't originally name, I went and learned the names for.
Fewer can ship the system that runs it.
Try out the demo here