🍉 CyberGym
Cybersecurity from zero, AI-Security beyond. A public, lifelong cybersecurity curriculum, written from the ground up.
What this is
CyberGym is a public knowledge base structured as a lifelong personal curriculum for cybersecurity, with a particular focus on the Cloud Security and AI/ML Security crossover.
It is not a journal. It is not a tutorial collection. It is a course — structured into phases, with numbered chapters and lessons, each one written to take a reader from no prior knowledge to working understanding.
Every note is written in plain English at street-smart adult, non-CSE level. Every term is defined the first time it appears. Analogies first, technical detail second. No theory walls. No condescension.
Who this is for
- Anyone with no computer-science background who wants to understand how a machine works before learning to defend it
- Anyone who has been told “just read the OSI model” and felt patronised
- Anyone curious about where AI security goes from here
How the vault is built — the layered-cake model
Most cybersecurity curricula go deep on one topic before touching the next. By the time the reader finishes 12 weeks on hardware, they have forgotten why they started.
CyberGym uses a spiral curriculum instead — six topic areas visited at growing depth across four phases. Like building a 100-story building: the skeleton goes up across all 100 floors before a single bag of cement is poured.
graph LR P0[Phase 0<br/>Foundation<br/>kindergarten → masters] --> P1[Phase 1<br/>Security+] P1 --> P2[Phase 2<br/>AWS Security Specialty<br/>Cloud Security niche] P2 --> P3[Phase 3<br/>GIAC GAIA<br/>AI/ML Security endgame] P3 --> P4[Phase 4<br/>Practitioner / portfolio]
The same six topic areas appear in every phase, getting deeper each time:
- Machine — hardware, boot, virtualization, GPU
- OS — Windows, Linux, containers, runtimes
- Networking — LAN → cloud → ML traffic
- Security — vocabulary → defender practitioner → AI threat modelling
- Cloud — intro → AWS hands-on → security specialty → AI workload security
- AI/ML Security — vocabulary → OWASP LLM Top 10 → labs → GAIA-level
Depth ceiling
Each topic stops at “a working defender on a security team can explain this and use it in an investigation.” That is working-defender depth. The vault does not descend into transistor physics, branch-prediction internals, or research-paper depth — those are research careers, not what gets people hired in cloud and AI security.
Where to start reading
- 00-Course-Map — the whole curriculum as a tree. The map.
- Welcome — the very first note. Start here from zero.
- What is a computer, really? — first real lesson.
Status
- Current phase: 0 (Foundation)
- Currently writing: Chapter 0.0 (Start Here) and Chapter 0.1 (Physical Machine)
- Stub-only chapters: 0.2 through 0.7 — content arrives as the curriculum is built out
- Started: 2026-04-30
How notes work
Every note here is a living document. It will be re-read, edited, rewritten in places it didn’t quite land, expanded where deeper detail is wanted. If a note feels rough, that is because it is v0.1 — it will be sharper next pass. The whole point of the spiral is that each topic is revisited at a deeper level later.
To follow along: bookmark this page, or open the course map and dive in wherever interests you.
Built with Obsidian, Quartz v4, and stubbornness.