If I claim one skill, it’s troubleshooting across layers. Phones, kernels, networks, clouds, build pipelines, my own gadget pile. The method is always the same: get the system to show you what it’s actually doing, find where reality diverges from the documentation, fix it or write it down, then teach it.
The stack below is the map of the work. Each layer had its years, and the receipts for every box live in the timeline.
Diagram source (mermaid)
%%{init: {"theme":"neutral", "themeVariables":{"fontSize":"16px"}}}%%
flowchart TB
A["Hardware and Gadgets<br/>always: home lab, gadget notes, RPi"]
B["Linux and OS<br/>2000s: server admin roots"]
C["Mobile and Custom ROMs<br/>2011 to 2015: Tamer Platform, ROM security"]
D["Network and Web<br/>2013 to 2019: pentesting, Python tooling"]
E["Cloud and DevSecOps<br/>2019 to 2022: IAM tooling, cloud training"]
F["CI/CD Pipelines<br/>2021 to 2023: DevSecOps research and training"]
G["Software Supply Chain<br/>2023 to now: SBOM Play, KeyChecker, supply chain training"]
H["AI Security<br/>2025 to now: AIDC, AI BOM Maker, private AI infrastructure"]
A --> B --> C --> D --> E --> F --> G --> H
classDef layer fill:#f4f4f4,stroke:#555,color:#111
class A,B,C,D,E,F,G,H layerThe subjects changed with the field: custom ROM research in 2011, SBOMs in 2026. The work underneath stayed the same. When AI arrived I moved one more layer up and asked the same question there: what is this system actually doing? That question became AIDC and the AI BOM Maker.
I wrote the longer argument in Troubleshooting: The Survival Skill We Forgot We Needed. The points I keep returning to across all of this are collected on Common Points.