If a learner struggles, I ask what we designed badly first.
Six commitments run under every project I ship — grounded in learning science, tested against the hardest environment I design for: incarceration and reentry. The same commitments hold anywhere the stakes are real and the margin for a confusing screen is zero.
The method underneath all six: a friction audit— every barrier that stops a specific learner, split into what's environmental (the device, the room, the interruptions) and what's internal (working memory, motivation, what a losing streak does to whether trying still feels worth it). Diagnose the friction, then build the fix — fast enough to prove it works, not just propose it.
Learning designed to be experienced, not endured.
Start from the real stakes
Design backward from the moment the skill actually gets used — the interview, the job center — not the syllabus.
Prototype, don't propose
Build it playable and iterate (SAM / LXD), instead of handing over a document and hoping.
Grounded in learning science
Cognitive load, retrieval practice, Merrill's First Principles, andragogy — why it works, not just what looks good.
Aesthetics are instruction
How a screen feels decides whether a learner will risk a mistake. Craft is pedagogy, not decoration.
Designed for dignity & access
Trauma-informed pacing, UDL, and the real environment — locked devices, no internet — as the starting point.
Made with AI, judged by a human
AI accelerates the build; the call on what's accurate and worth teaching stays mine.
Every “Grounded in” tag on a case study points back here.
Foundational Models
- ADDIE
- SAM
- Backward Design
- Merrill's First Principles
Learning Theory
- Andragogy
- Constructivism
- Cognitive Load
- Scaffolding
- Multimodal
Inclusive & Trauma-Informed Design
- Trauma-Informed
- UDL
- Accessibility (WCAG)
- Learner Autonomy
Evaluation
- Kirkpatrick
AI is my force-multiplier — I direct the instructional design; AI accelerates the build.
Trained LLMs, from the inside
At Character.AI I wrote and evaluated thousands of dialogue examples alongside ML engineers — I know these models as materials, not magic.
Built my own AI tools
A RAG instructional-design assistant (LangChain, ChromaDB) that answers design questions against my own source library, with citations.
Built this course with AI
The Digital Literacy course and its live simulation — I directed the design; AI accelerated the build from storyboard to prototype.
AI-native production
Prompt-chained drafting, structuring, and quality-checking — more output, the same judgment over what's accurate and worth teaching.
See these six commitments applied, not just stated.
See the work