Hello, I'm
Security & infrastructure engineer working across hardware, software, and systems.
I'm a jack of all trades who has worked across the stack — from ASIC and FPGA hardware design to software development, DevOps, and sysadmin work. These days I focus primarily on information security.
I believe in understanding systems from the silicon up. That breadth is not a lack of focus — it's the ability to see threats and solutions that specialists miss. A vulnerability in a hardware design looks different when you've written RTL. An infrastructure problem looks different when you've built the pipeline that deploys it.
Before the LLM era, I worked on neural networks and studied the underlying mathematics. The field has changed, but the foundations haven't. I'm interested in where security and AI intersect — adversarial systems, prompt injection resistance, and building autonomous agents that can be trusted.
Outside of work, I run a homelab that I've rebuilt more times than I can count. Each iteration was an excuse to learn something new. The current incarnation is i.ar — an infrastructure-as-code project that ties together WireGuard mesh networking, containerized AI agents, and a hardened Emacs environment running on local hardware.
ASIC and FPGA development. RTL design, verification, and the low-level thinking that comes from building things that can't be patched after tapeout.
From embedded firmware to infrastructure automation. Python, Bash, Emacs Lisp, and whatever tool fits the problem.
Ansible, Podman, WireGuard, Caddy. Infrastructure as code with reproducible deployments. Immutable OS foundations with SELinux.
Container hardening, attack surface reduction, prompt injection resistance. Building systems that assume they will be attacked.
Pre-LLM neural network experience. Now building agentic systems with local models, focused on security and trust.
Linux systems, Fedora Silverblue, SELinux, systemd. The unglamorous work that keeps everything else running.
A self-modifying AI operating environment built in Emacs, hardened in Podman, and powered by local LLMs. No cloud. No telemetry. No backdoors.
The project ties together a WireGuard mesh network across three cloud servers and two local machines, a containerized agentic Emacs workspace with multi-agent delegation, and infrastructure-as-code deployment via Ansible. It started as a homelab — rebuilt more times than I can count — and evolved into a platform for exploring where security and AI intersect.
I'm open to opportunities in information security, infrastructure, and AI. If you're building something interesting, I'd like to hear about it.