Cloud, on-prem, or a bit of both. I design it, build it and keep it running for businesses that really can’t afford downtime. Oracle Cloud, AWS, GCP, VMware, Proxmox, Kubernetes, the lot. I use AI to move faster, but every change still gets checked by me.
3 nodes, 6 services, all happy. Go on, click a node and break it.
I’m Niloy, an infra and cloud engineer based in Glasgow. I started out in enterprise support as the escalation guy for VMware and HPE platforms, the person you get once the first two lines have run out of ideas. That’s where I learned how big systems actually break. Since 2023 I’ve been doing my own thing through TwentyOneLabs, designing and running cloud and on-prem platforms for businesses. Oracle Cloud, AWS, GCP, VMware, Proxmox, Kubernetes, whatever fits the job.
Senior infrastructure & cloud engineer
Glasgow, UK · working remotely
Short version: years of fixing other people’s biggest outages, then I started building my own stuff. Tap a role for the details.
From picking the right cloud to the 2am patch window. I build it, write it down, and hand it over properly.
Moving to Oracle Cloud, AWS or GCP, or tidying up what you’ve already got there. Sized properly so the bill doesn’t surprise you.
VMware, Proxmox, Hyper-V or KVM clusters that keep going when a host dies. Including getting off VMware if the licence bill has gone silly.
Docker and Kubernetes with CI/CD and monitoring. Right-sized, so no 40 node cluster for a blog.
Segmented networks, locked down access and an edge that only lets the right people in.
DR plans that get tested, not just written. Clear recovery targets, off-site copies and restore drills.
Hooking software up to real-world equipment, and scripting away the boring, repetitive jobs.
I use AI every day. It drafts the code, the docs and the first go at a fix. Then I review it, try it in a lab, and only ship what I’d happily put my name on.
Cloud and on-prem, I’m at home in both. Tap any card to flip it and see how I’ve actually used that tool.
8 categories · 36 tools · 7 legendary
Where most of my recent work lives. Oracle Cloud is home base, with AWS and GCP right behind it.
Same four steps every time. Boring? A bit. But boring is exactly what you want from infrastructure.
I look at what you’ve got, what’s hurting and what you actually need. No money spent yet.
A written plan with costs, trade-offs and a way back if things go sideways. I try it in a lab first, so the surprises happen there.
Infrastructure as code, staged cutovers and tested backups. Changes go through review and a proper maintenance window.
Monitoring, patching and firmware under change control. I either hand it all over with docs, or stick around on a retainer.
Freelance work through TwentyOneLabs. From a national motorcycle racing team to the back office of a UK recycling company.
Hiring, a project, or just a question about your setup. Send a few lines and I’ll get back to you, usually within a working day.