Jack Devlin · Northern Ireland · platform engineer
available from October 2026
Hey, I'm Jack
Platform engineer, based in Northern Ireland. Day job is Clarity, a natural-language database product running across about thirty tenants, and the LLM gateway every AI workload at the company goes through. I built the Kubernetes, pipelines and observability underneath them as well. I'm finishing an MSc in Artificial Intelligence in September 2026.
Where I've worked
Named, dated and checkable. There's a PDF of the same thing if you'd rather have one.
Loweconex — a UK IoT platform business
Aug 2023 – present
Platform & Site Reliability Engineer · Lisburn, Northern Ireland
Core platform engineering, on-call and AI infrastructure for a high-volume microservices estate. Engineering was five people when I joined and is around forty now, so most of what I built was built to keep up with that. I started as a graduate in QA, moved into platform work, and have been contracting through my own company since September 2025.
What came out of it: Clarity, AI Gateway, Heimdall, Pipeline platform, Observability stack.
OD3 Engineering
Aug 2021 – Aug 2022
Software Developer (industrial placement) · Magherafelt, Northern Ireland
A placement year building Tekla Structures API applications and design-platform integrations, automating drafting work for the in-house architects.
Education
MSc Artificial Intelligence, Distinction
completing September 2026
Queen's University Belfast
Dissertation: capacity-aware recovery scheduling for Kubernetes, measured on real EKS clusters under induced node failure.
BSc Computer Science, 2:1
2019 – 2023
Queen's University Belfast
First-class dissertation on deep learning for stock-price prediction.
Since September 2025 I've worked through Devlinops Ltd, my own limited company — devlinops.com is a portfolio, not an agency. Irish and British citizen. Full right to live and work anywhere in the EU as well as the UK, no sponsorship required.
How it went
Three years at Loweconex, and the scope kept growing. Each year built on what the last one shipped.
2021
Placement year at OD3
A year out of my degree writing Tekla Structures API applications, automating drafting work for the in-house architects. First time I'd shipped anything people depended on.
2023
Joined as a graduate, inherited a migration
I finished my BSc and landed in the middle of a monolith → microservices migration, writing the test automation that kept it honest. Engineering was five people.
2024
Earned the platform
The test work kept exposing infrastructure problems, so I started fixing those instead. Observability from zero, then GitOps with ArgoCD, until 20 services deployed the same way.
2025
Standardised it, then instrumented it
I moved CI onto one shared pipeline library across every service, and started the deployment-metrics tool that grew into Heimdall. The team went from asking "did it deploy?" to reading the answer off a screen.
2026
Built the AI platform on top
The same infrastructure, now carrying LLM workloads: a gateway in front of every model call, and Clarity, a natural-language database product live across ~30 tenants. I'm finishing an MSc in Artificial Intelligence alongside it.
principles
How I work
Three things I've come to believe after a few years on platform teams.
A correct system nobody opens is not finished
The DORA collector behind Heimdall was right for months and nobody ever looked at it. Same data, no front door. Building the UI is what turned it into something twenty people use every morning, and I've stopped treating the interface as the optional half.
Boring is a compliment
The ideal platform fades into the background. Engineers shouldn't have to think about it any more than they think about the office wifi.
Operability is a feature
If a teammate can't tell whether your service is healthy in under a minute, you haven't finished it yet. I default to one-curl health checks and a runbook per alert.
stack
Tech stack
Split three ways, so you know which is which: what I've been on call for, what runs in my flat, and what I've only used.
run-in-production/
Evidenced by the case studies. I have carried a pager for these.
homelab/
Running in my flat, reconciled the same way as work. Nobody pages me, but it does break.
working-knowledge/
Used on real projects or in coursework. I haven't been on call for any of these.
Right now
Shipping Clarity · Finishing the MSc · Available from October 2026
I'm available from October 2026, permanent or contract, for platform engineering, developer experience, observability or AI infrastructure work. I'm looking at Dublin, London and Amsterdam, or remote-first anywhere in the EU — and as an Irish and British citizen I need no sponsorship for any of them.
I'm wrapping up my current contract on the platform team I helped build. Most of this year went to the AI side: Clarity live across ~30 tenant databases, and the gateway that fronts every AI workload. Heimdall still opens every morning.
The MSc finishes in September 2026. The dissertation builds a capacity-aware scheduler for recovering Kubernetes workloads after node failure, measured on real clusters where failure means actually killing the machine. It's the same problem I keep hitting on the platform side, so the two halves have converged.
Still reading?
Drop me a note. About a role, an AI infrastructure problem, or anything that overlaps with the work above.