A learning platform for subjects that are difficult, deep, poorly taught, or require digging through specifications, source code, academic papers, and fragmented documentation.
Underlayer takes things that normally require enormous amounts of scattered research and turns them into structured, interactive, deeply understood knowledge.
Courses:
- ELF — Executable and Linkable Format — the technical deep dive the platform started with.
- HAT — Higher Education Aptitude Test — preparation for the HEC Higher Education Aptitude Test, written for HAT-1 candidates in engineering, computing and the physical sciences (5 modules, 21 concepts).
- Courses are long-lived artifacts, improved continuously for years
- Optimizes for understanding and retention, not course completion
- Every interaction has pedagogical purpose
- Mistakes are learning opportunities, not failures
- Difficulty is evidence of learning, not a sign of incompetence
Learners who:
- Struggle with anxiety, depression, or overthinking
- Need repetition and slow, consistent pacing
- Want concepts hard-wired into their brains, not surface-level recognition
- Are tired of scattered blog posts, 400-page specs, and 12-hour video courses
- Want to actually understand how things work, not just pass a quiz
Chemical — a systems programming language.
Free and open source. Courses are portable, self-contained artifacts.
Planning phase. No implementation yet.
AGENTS.md— Rules for AI agents working on this projectdocs/conceptual-model.md— Data model for courses, concepts, learner statedocs/plan.md— 6-phase implementation roadmap
docs/course-development-handbook.md— Step-by-step 7-phase guide for AI to develop coursesdocs/ai-course-writing-constraints.md— 8 constraint methods for AI generation qualitydocs/course-design.md— How courses are structured for deep learningdocs/features.md— 10 core features + 6 advanced features
docs/teaching-components-catalog.md— Complete catalog of all teaching primitivesdocs/rendering-pipeline.md— How .ch files become interactive HTMLdocs/developable-components.md— What AI can generate vs what requires engineering
docs/ui-ux-design.md— Colors, typography, spacing, components, accessibilitydocs/competitors.md— Brilliant, Exercism, CodeCrafters, roadmap.sh, nand2tetris, OST2, Coursera
docs/learner-profiling.md— IRT diagnostic, behavioral profiling, adaptation rulesdocs/adaptive-flow-ui.md— Screen-by-screen flow, knowledge display, anxiety/depression design
docs/correctness-verification.md— 5-layer verification system against standardsdocs/deployment.md— Android app and auto-deployment.agents/skills/— 10 AI skills for working on Underlayer