Human Engineering
Write everything yourself. Understand every layer.
- Full control over every decision
- Clear reasoning behind every line
- Built for deep understanding
Learning Rust theory but not building anything? Step into real engineering systems. Assemble, modify, and break them locally — guided, but you build it.
Pick your playground →
You’ve seen it all — books, tutorials, videos, articles, code snippets, toy problems, tasks without context. All you really needed was to start writing code.
Start building →This is not passive learning. You build, with guidance, feedback, and real code.
Unlock the Full Experience
Work locally, make changes, and take ownership over time — with guidance when you need it.
Start with guidance. Progress to full ownership. Hints fade. Decisions become yours.
No cloud. No lock-in. The code is yours — on your machine, forever.
No simulations. No cloud sandboxes.
Build and run everything on your laptop.
Install
Install the local toolchain first, then connect your own AI agent, such as Claude Code, Codex, Gemini CLI, or similar. You will need an agent.
cargo install knowledge
You assemble components, explore trade-offs, and debug failure modes — with guidance, reference solutions, and responsibility.
You start with guided assembly and full reference solutions to understand the system deeply.
Later, guidance steps back — and responsibility moves to you.
Use VS Code, Vim, or any editor you prefer.
Working in your real development environment builds the same workflows and habits you'll use in practice.
Use Claude Code, Codex, Gemini CLI, or another agent you already trust.
The playground fits into your real workflow instead of replacing it.
The reference project is available on request when you need it.
Compare your code with the reference implementation, inspect differences, and learn by reconstructing it yourself.
See exactly what changes at each step.
Every modification is shown as a clear diff, with explanations for why it matters.
Every playground leads to a finished application you can run, inspect, and extend locally.
Not fragments, not demos — complete systems with structure, state, and behavior.
Ready-to-use prompts for AI coding assistants.
Recreate each step using Claude, Cursor, or your preferred tool — if you want shortcuts.
Your AI agent can communicate with you in your native language.
Any language is supported.
Every playground is yours to explore, modify, break, rebuild, and extend.
You are free to move through the project your own way and make it truly yours.
To get started, connect your AI agent to the local Knowledge.Dev MCP server, choose a playground, and start building.
Your progress, course state, and reference files stay within the MCP workflow, while the project runs locally on your machine.
Enter playground →
Why choose?
Both.
Build with your own hands when you want full understanding — or delegate implementation to AI while you stay focused on architecture and outcomes.
Write everything yourself. Understand every layer.
Let AI implement while you stay in control.
Engineering notes, new playgrounds, and platform evolution.
Read the blog →
