I work on the systems underneath the product.

I have spent much of my career working on the systems underneath the product: infrastructure, deployment, security, observability and production operations.

I've worked in very small startups, scaling SaaS companies and enterprise environments, and I've seen the same pattern repeatedly: infrastructure grows faster than the operating model around it. Accounts, services and pipelines multiply. The decisions about who owns what, and how it should all fit together, lag behind.

My job is usually to help bring that complexity back under control.

I started as one of the first DevOps engineers at a small SaaS startup, and stayed through the kind of growth that teaches you what breaks at each stage, through to its acquisition. Since then I've designed and operated AWS infrastructure spanning more than 50 accounts across multiple regions, and built highly available platforms for SaaS, real-time and AI workloads. I've held senior technical leadership roles up to and including CTO.

Today my work spans security, cloud operations, platform engineering and production operations. Security and compliance are a growing part of it: helping companies meet ISO 27001, SOC 2 and enterprise customer requirements with controls that actually hold. More of my work is also AI infrastructure: GPU workloads, model serving with vLLM, QLoRA fine-tuning, and the question of whether a given workload belongs on AWS or GCP.

I'm still hands-on. I enjoy the leadership conversations about how production should be run. I also enjoy opening a terminal and finding out why it isn't running that way. I think that combination is the most useful thing I offer. Advice is better when it comes from someone who will also implement it, and implementation is better when it comes from someone who understands why the business needs it.

What I care about

• Boring is a feature. Production should be predictable. Clever infrastructure is usually just infrastructure that's hard to hand over.
• Simplify before adding. The first improvement is often removing something.
• Security as engineering. Controls that the platform enforces beat controls that people have to remember.
• Ownership over heroics. A system is healthy when your team can run it without me, or anyone else in particular.