abukix

John Carl Abucay

SRE · building an AI platform in public

kernel to LLM

stillness between commits.

Months

Phases

Projects

✦ Built in Public

Building an AI platform in public.

60 months. 30 phases. 9 open-source projects. Patterns over tools — find the timeless principle first, pick a tool to master second.

Copy $ git clone github.com/abukix/portfolio $ # ...60 months of deliberate practice $ # one homelab. one platform. built one phase at a time.

Explore the plan, the projects, and the homelab inside.

GitHub LinkedIn YouTube Email

Building an AI Platform in Public

60 months. 30 phases. 9 open-source projects. Patterns over tools.

stillness between commits.

⌘ K

Projects

9 open-source projects across the 5-year plan.

Explore →

Skills

Technologies and tools I work with.

Explore →

Architecture

The 5-year plan, in tiers.

Explore →

Experience

Roles, programs, and milestones along the way.

Explore →

Certifications

Validated credentials and exams passed.

Explore →

Blog

Writing and thinking out loud.

Read →

Connect

Ways to reach me or follow along.

Explore →

Homelab

Hardware + roadmap for the platform on my desk.

Explore →

Resume

One-page overview.

View →

stillness between commits.

Projects

9 open-source projects across the 5-year plan

Name	Type	First Touched	Description
ops-handbook	docs	Year 1 Phase 1	Runbooks, incidents, postmortems — runbook-as-code
rxp	CLI	Year 1	Text processing — programming patterns
konfig	CLI	Year 1	Schema validation
pulse	CLI	Year 1	Concurrent fan-out scanner
triage	Service	Year 1-2	First real K3s service
basecamp	GitOps	Year 2 → 5	Declarative platform · grows through Year 5
terralabs	IaC	Year 2 → 5	Declarative infrastructure modules
platform-ctl	CLI	Year 2 (Phase 12)	Platform CLI — platform-as-product
mlship	PyPI	Year 5 capstone	One-command ML model deployment

Status: scaffolded. Each ships publicly as its phase begins.

Skills

Tools the platform end-state uses, organized by tier

Core

In use

Linux Python Bash Proxmox

Containers + Orchestration

Year 1

GoDockerKubernetesHelm

Infrastructure

Year 2

TerraformArgoCDAWSGCP

Data

Year 3

SparkAirflowTrinoIceberg

ML + AI

Year 4

MLflowRayKubeflowvLLMRAG

Observability

Year 3

PrometheusGrafanaOpenTelemetryeBPF

Agents + AI Platform

Year 5

LangGraphMCPAIOpsKeycloak

Skill bars reflect the phase they land — not self-rated expertise. Bars fill as each phase ships.

Architecture

The 5-year plan, in tiers

AI Platform + Capstone

Year 5

LangGraph MCP AIOps Keycloak Portal mlship

ML & AI Infrastructure

Year 4

MLflowRayKServevLLMKubeflowFeast

Platform Engineering & Data

Year 3

PrometheusGrafanaeBPFSparkAirflowIcebergTrino

Distributed Systems & Cloud

Year 2

TerraformArgoCDAWSGCPCrossplaneBackstage

Foundations of Computing

Year 1

LinuxNetworkingDatabasesPythonGoDockerKubernetes

Foundation

Distributed

Data

AI Platform

Experience

Roles, programs, and milestones along the way

2023 — present

SRE → AI Platform Engineer

At a major tech company

Supporting and building data/AI infrastructure. Working with Kubernetes, distributed systems, and cloud-native platform services.

2026 — ongoing

ROOT — 60-month program

Self-directed · public

A 60-month, 30-phase program teaching engineering by finding root principles, not chasing tools. Currently Year 1 Phase 1 — OS foundations. Documented at abukix.dev/blog.

2024 — 2026

Certifications

5 industry credentials

CKA, LFCS, KCNA, KCSA, AWS Cloud Practitioner. Validated through hands-on performance-based exams.

Certifications

Validated credentials and exams passed

CKA

Certified Kubernetes Administrator

Linux Foundation

2024

LFCS

Linux Foundation Certified Sysadmin

Linux Foundation

2025

KCNA

Kubernetes Cloud Native Associate

Linux Foundation

2026

KCSA

Kubernetes Security Associate

Linux Foundation

2026

AWS CP

Cloud Practitioner

Amazon Web Services

2025

Connect

Ways to reach me or follow along

Homelab

Hardware + roadmap for the platform on my desk

ThinkCentre M70q Gen 5

16GB DDR5 256GB NVMe 16 threads Proxmox

Bastion VM

Ubuntu 24.04 key-only SSH snapshot baseline

Coming Year 1-2

K3s (Phase 7) Multi-cloud (Year 2) 32GB upgrade Mo. 10

The Planned Stack

ArgoCD

Prometheus

Grafana

PostgreSQL

MinIO

Spark

Airflow

Redpanda

Kubeflow

MLflow

KServe

vLLM

Trino

Superset

Ray

Feast

Keycloak

Cilium

Sealed Secrets

Velero

The end-state stack — built progressively across 5 years. Each service lands as its phase begins.

Resume

One-page overview

JC Abucay

SRE · building an AI platform in public

hello@abukix.dev · github.com/abukix · linkedin.com/in/abukix

Experience

SRE → AI Platform Engineer 2023 — present

At a major tech company. Supporting and building data/AI infrastructure.

Platform (in progress)

Building a homelab-scale data + AI platform across 60 months — same architecture patterns as production. 30 phases, 9 open-source projects. Currently Year 1 Phase 1.

Skills

LinuxKubernetesTerraformArgoCDSparkAirflowKubeflowRayKServevLLMMLflowPythonGoAWSGCP

Certifications

CKA · LFCS · KCNA · KCSA · AWS Cloud Practitioner

Open Source

9 projects scaffolded across the 5-year plan: ops-handbook, rxp, konfig, pulse, triage, basecamp, terralabs, platform-ctl, mlship. Each ships publicly as its phase begins. See github.com/abukix.