Skip to main content
BrettOSv15.4
SYSTEM STATUS OPERATIONAL
HELLO, I'M

Brett Porter

Platform Architect · AI Infrastructure Builder

I build platforms that help engineers move faster.

SYSTEM STATUS OPERATIONAL
YEARS BUILDING SYSTEMS....0+
CLOUD SPEND REDUCED.......0%
UPTIME SUSTAINED..........0.00%
ENTERPRISE PLATFORMS......CNN · PGIM · Genentech
CURRENT FOCUS.............AI Infrastructure
STATUS.................... ONLINE
SCROLL
02Featured Architecture

Platforms, not job entries.

Each system below is a decision log — the problem, the architecture, the tradeoffs I made, the way it first broke, and what it taught me. Open one to read the case study.

PLT-01 · CCCTC

Multi-Cluster Observability

OPERATIONAL

Unified metrics, logs & traces across Kubernetes fleets for the largest public higher-education system in the United States — 116 colleges, ~2M students — policy-governed and self-healing.

ARCHITECTURE · SIMPLIFIED
Grafana LGTM
Prometheus
OpenTelemetry
Kyverno
Rancher 2
EKS
116
colleges, one fleet
2M+
students served
100%
policy-as-code
Grafana LGTMPrometheusOpenTelemetryKyvernoRancher 2EKSK3sTerraform+1
03The Control Plane

Step inside a real platform.

This is WhizAI — the AI infrastructure platform I built: an async workflow engine, a multi-provider model router with circuit breakers, and self-hosted vision, all on Kubernetes. Click any node to inspect it — open the cluster for the live run timeline and logs, or a service to watch requests flow.

API GatewayExpress · :4000 · authModel RouterFastAPI · circuit breakersWorkflow WorkerKafka consumer · step handlersWhizAI ClusterEKS · helm-base · 30 servicesQdrantvectors · per-projectKafkastep queue · DLQPostgreSQLPrisma · 116 modelsObservabilitymetrics · logs · tracing wip
Enter the control plane click any node to inspect
04The Journey

Fifteen years, told in eras.

Not a list of years — a progression of how the work changed. From milliseconds on air, to the cloud, to fleets, to platforms as products, to making AI economical. Open an era to step into it.

~2013Broadcast Systems Era
Vizrt · Oracle~2013

Broadcast Systems Era

Infrastructure for national broadcasters — Oracle databases and real-time graphics where downtime aired live.

REPRESENTATIVE WORK
Oracle DB — 10g/11g · RAC · ODA
On-air support: NBCU · Bloomberg · CBS
Broadcast workflow automation

Latency you can see on screen teaches respect for the hot path early.

~2015Cloud Era
AWS · Axial~2015

Cloud Era

Lifting systems into the cloud — elasticity, managed services, and the first real cost curves.

REPRESENTATIVE WORK
AWS Classic → VPC migration
Kafka · RDS · Redshift
Cost-aware scaling

The cloud removes the floor under your costs — and the ceiling. You own both now.

~2017Kubernetes Era
Containers · Orchestration~2017

Kubernetes Era

Standardizing compute on Kubernetes — containerization, operators, and the discipline of declarative infra.

REPRESENTATIVE WORK
Docker + Kubernetes adoption
Rook/Ceph storage · RBAC
Operators & controllers

Declarative infra is a contract. The cluster will hold you to it.

~2021Platform Engineering Era
Golden paths · DevEx~2021

Platform Engineering Era

Turning infrastructure into a product engineers want to use — golden paths, mesh, observability by default.

REPRESENTATIVE WORK
50+ service golden path
Istio multi-cluster mesh
Multi-cluster observability

A platform is a product. If engineers route around it, you built a maze, not a path.

2024+AI Infrastructure Era
Current focus · ACTIVE2024+

AI Infrastructure Era

Making AI workloads economical and reliable — inference engines, agent systems, and ruthless unit economics.

REPRESENTATIVE WORK
Multi-agent platforms
Cost-gated inference
RAG & vector infra at scale

AI doesn't change the job — it raises the stakes on cost, latency, and trust.

05AI Infrastructure Lab

Two tools, running live.

Not screenshots. Ask Brett answers architecture questions in real time, grounded in the systems above. The Architecture Critic reviews your Terraform or Kubernetes YAML like Brett would in a design review.

LIVE
BP

I'm Brett — or the part of me that reviews architecture. Ask how I'd build something. I'll give you the decision, the tradeoff, and the way it usually breaks.

"

The best architectures emerge from solving real problems at scale — not from diagrams on a whiteboard.

Brett Porter
LET'S BUILD SOMETHING THAT SCALES

Designing your platform layer or AI infrastructure?

SUMMARY

Platform architect, 15+ years. Broadcast systems → cloud → Kubernetes → platform engineering → AI infrastructure. I build the layer engineers build on.

SELECTED
  • Wunderkind — Platform Engineering
  • Code & Theory — CNN Datacloud / MagicWall
  • CCCTC — Multi-cluster Observability
  • Independent — Mean Scoop · MushAttention
BrettOS · v15.4 all systems operational© 2026 Brett Porter