Portrait of Matthew, aka TechDufus

You can call me Dufus.

Platform Engineer / Agentic Engineer

About

Hello, I'm TechDufus.

I'm Matthew, but most people online know me as TechDufus. I started in PowerShell, worked my way through DevOps, full-stack and platform engineering, and now I build with AI and teach others to.

Status: probably fighting DNS somewhere.

Portrait of Matthew, aka TechDufus
You can call me Dufus.Matthew · @TechDufusPlatform Engineer / Agentic Engineer
01Backstory

How I got here.

Nobody handed me a career path. Every time I ran out of new things to learn, I found the area I had the least exposure to and switched into it.

  1. Where it started PowerShell automation

    Got bit by the automation bug in a heavily regulated shop. Built AdminToolkit and HelpDesk, stopped clicking through runbooks, and picked up Security+.

  2. Wanted the whole pipeline DevOps

    Scripts only covered my corner. I wanted to see how software actually ships, so I moved into DevOps: Terraform on Azure, CI/CD, PKI automation that went from hours to seconds. CISSP along the way.

  3. Wanted the app side too Full-stack

    I wanted to understand the apps I was deploying, so I joined a team that owned everything from app code to firewalls. Go services, Helm on GKE. Learned to care about the people using what I build.

  4. Wanted the platform underneath Platform engineering

    Big Kubernetes clusters, RBAC at scale, GitOps with ArgoCD, air-gapped environments, Ansible across whole fleets. Fast only counts if you can roll it back.

  5. Now AI everything engineer · Team lead, solutions architect

    AI means one person can finally work across the whole stack, so I do: infra, code, and the agents that write it. I lead a team now, and most of that job is mentoring and training people to use AI well. I'm also building YahwAI, an AI Bible-study tool, on the side.

If I repeat something twice, I script it on the third pass.

02NowUpdated February 26, 2026

Right now.

AI inference and GPU workloads on Kubernetes, and pushing Codex and Claude Code until the rough edges show.

03Proof

Proof.

Production Scale

200-node AKS + 100TB+ data platform[1]

The architecture is written up publicly.

Secure Delivery

High-security + air-gapped platform operations[2]

GitOps and compliance in high-stakes environments.

Agentic Work

Multi-agent workflows with human ownership

Fast iteration with explicit review and rollback.

  1. Secure Low-Latency Queries at Scale with Raft Data Platform — teamraft.com ↗
  2. Related coverage: Fighter pilots take directions from AI in Pentagon's groundbreaking test — foxnews.com ↗
04Principles

How I work.

  • 01

    Reproducible by default

    If Terraform can rebuild it from a fresh box, I trust it.

  • 02

    Automate early

    One-off steps drift fast.

  • 03

    Fast, with guardrails

    Risky changes still get review and a rollback plan.

  • 04

    Write it down

    If it only lives in someone's head, it's a risk.

Say hey!

Seriously, please talk to me.

Open the contact page