JC Abucay
SRE becoming an AI/ML platform engineer, building in public.
I work Support/SRE on an internal, enterprise-scale data platform. Alerts, triage, production incidents: the operational side of a system someone else designed. That has been most of my career so far.
A while back I started wondering what it would take to be the one designing the system instead of the one paged for it. Not a bootcamp. A real curriculum, written by me: five arcs, fifty phases, starting at Linux fundamentals and ending at ML platform infrastructure. I call it /root.
Tools change every five years.
Patterns don't change in thirty.
That is the bet /root is built on. Every phase names the underlying pattern before it names the tool, because tools get replaced and patterns do not. The curriculum is fully authored now: fifty phases, five arcs, all of it public.
The platform the curriculum is meant to produce is basecamp, an open-source data, ML, and AI stack meant to run the same way on a laptop, a homelab, or the cloud. Right now it is architecture and a roadmap. No module code has shipped. I would rather say that plainly than let the repo imply otherwise.
Selected work
- /root: The curriculum. 50 phases across 5 arcs, fully authored.
- basecamp: The platform. Architecture and roadmap only, no module code yet.
- bootstrap: Idempotent dev-laptop setup for Ubuntu, Debian, and macOS.
- learnix: A learning-in-public archive of cloud and platform engineering notes.
- brand: Design tokens and voice guidelines behind every abukix artifact.
- homelab: Hardware and OSS stack notes for the homelab running basecamp.
Support the journey
None of this is finished, and that is the point of building it in public. If a repo above is useful to you, a star helps more than it might seem: it is the signal that tells me which parts are actually worth finishing first. A follow on GitHub means you see the next one when it ships.
Reach me
Questions, corrections, or you're working on something similar and want to compare notes: any of these reach me.
Everything above is public because half-finished, in-progress work is still worth showing honestly.