Open to work — relocating welcome, Dublin especially
I'm a cloud and platform engineer. I like infrastructure that's reproducible and quiet — the kind you forget is running until it quietly recovers on its own.
I'm finishing a two-year apprenticeship as an SRE/DevOps engineer at Enedis, alongside my engineering degree. I'm looking for a full-time platform or cloud role from September 2026.
Featured project
Codir is a side project I took further than I planned: an enterprise-management SaaS I run on my own Kubernetes cluster, delivered end to end through GitOps. It's event-driven — services publish to RabbitMQ instead of calling each other directly, so each one deploys and fails on its own without dragging the others down. The diagram below is the real architecture; tap a component to see what it does and the call I made.
More work
A mix of personal builds and paid work. The personal ones I can show in depth; client and employer work I describe, since the code isn't mine to publish.
A single Proxmox host — 8 cores, 64 GB — that I run like production. Around 20 VMs and LXC containers stay up around the clock; five or six are reachable from outside on demand through a Cloudflare Tunnel, so there are no open ports and nothing permanently public. One of them runs Frigate for real-time AI object detection on my camera feeds. It's where I test the patterns I later trust at work: IaC, segmented networking, backups, monitoring.
A multi-user platform I built and operated on AWS (ECS Fargate, API Gateway, Lambda, RDS), provisioning isolated environments on demand, with Terraform IaC and self-healing incident remediation. It's onboarded several teams so far, and it's built to scale to roughly 100 project teams. Described here — the code is private.
An accounting firm was tracking hours across scattered Excel files that kept drifting out of sync. I built them a FastAPI + React app that became the single source of truth — one validated format, no more hand-reconciling spreadsheets. That ended the tracking errors those files bred and gave the firm a single view of its hours, in a shape that meets their legal reporting duties. Shipped to production, with a 20-scenario acceptance suite passing 100%.
A browser-based traffic simulation built on the IDM and MOBIL models. I wanted to see emergent behaviour first-hand — how phantom traffic jams form from nothing — and watch flow and density react in real time.
Capabilities
Track record
I scoped the platform's architecture with the SRE team, then ran it in production on AWS — fully provisioned as code, with incident auto-remediation and a TTL cleanup that kept orphaned environments and cost in check.
End-to-end test suites and automated CI/CD pipelines.
High-capacity virtual desktop infrastructure (VDI) project.
Infrastructure, web apps and e-commerce for small businesses.