role # senior platform infrastructure engineer

I keep infrastructure boring —
so the product can be interesting.

Platform engineering, Kubernetes, and observability — currently retiring a legacy auth system service-by-service and building toward AI infrastructure: RAG pipelines, vector databases, and the platforms that keep them reliable.

status — operational
focuslegacy auth sunset
last_deployauth gateway ✓
learningrag / vector dbs
availabilityopen to contracts
locationremote-first 🌍
selected_work

Projects & case studies

portfolio_project / ongoing

Autonomous DevOps Reliability Platform

A self-healing reliability platform built on k3s: full-stack observability with Prometheus, Grafana, and OpenTelemetry, wired to GitHub Actions for automated remediation and deploy pipelines. Designed to answer one question — how much of on-call can a platform absorb before a human gets paged?

k3sprometheusgrafanaopentelemetrygithub actions
read the case study
case_study / production

Sunsetting a legacy auth system

Decommissioning a company-wide legacy auth service, one dependent service at a time — most recently the auth gateway itself. Scope, migration strategy, and how to retire critical infrastructure without anyone noticing.

migrationauthzero-downtime
read the case study
exploration / ai_infra

RAG on my own metal

Building retrieval-augmented generation pipelines and evaluating vector databases on local hardware — treating AI systems as an infrastructure problem: latency, reliability, cost, and observability.

ragvector dbslocal llms
notes coming soon
writing

Blog

2026 · draft How we sunset our legacy auth system without breaking production coming soon planned What platform engineers should know before touching RAG coming soon
homelab # infra for fun, not just for a paycheck

The lab

mini-pc · ubuntu server

Always-on core of the lab. Runs self-hosted services and acts as the testbed for anything before it earns a place in a real cluster.

tailscale

Mesh VPN tying every device together — the lab follows me anywhere, which matters when "anywhere" is the plan.

immich

Self-hosted photo management. Because a platform engineer's photos shouldn't live on someone else's platform.

linux workstation · rtx 5070

Local compute for AI work — running and evaluating local models, embedding pipelines, and GPU workloads without cloud bills.

about

Hello

I'm a senior platform infrastructure engineer. My day job right now is a long, careful demolition: retiring a legacy authentication system service-by-service, most recently the auth gateway. It's the kind of work where success means nobody notices anything happened.

I'm betting the next few years of my career on the overlap between platform engineering and AI infrastructure — the unglamorous layer that makes RAG systems, vector databases, and model serving actually reliable in production. I'm building that expertise in public, partly on my own hardware.

I'm working toward independent contract work and a location-flexible life. When I'm not staring at dashboards you'll find me traveling, surfing, or on a dance floor somewhere.

off_hours:
  • travel — slowly, one region at a time
  • surfing — perpetually intermediate
  • social dancing
contact # open to platform & ai-infra contracts

Need someone to make your infrastructure boring?