Senior Site Reliability Engineer

Department: Engineering

Employment Type: Full Time

Location: London, UK

Compensation: £95,000 - £105,000 / year

Steel Eye, now merged with Fund Apps, provides communications, trade and order surveillance to some of the world's largest financial institutions. Clients run on shared or dedicated infrastructure depending on what they need, which means we operate a large estate of clusters, data pipelines and client environments for the size of our engineering team.

This role keeps that estate healthy. Not just up, but doing its job: client data arriving completely, surveillance running against the right window, the alerts our clients depend on actually firing. When something is wrong we want to know before they do.

It is a hands‑on engineering role. You will spend your time in Terraform, Kubernetes manifests and code rather than in meetings about them.

Running the Kubernetes estate. We run many client environments across EKS. You will keep them properly separated, sensibly sized and able to scale when a client sends us more data than we expected. Isolation between clients is a client assurance commitment, not just good practice, so the boundaries matter.

Knowing what is happening. We have a substantial monitoring and alerting setup and it needs to keep pace with the platform. You will extend it, keep the signal useful, and push ownership of alerts and service levels out to the teams that build the services rather than holding it all centrally.

Making deployment boring. Changes should reach production frequently, in small pieces, without drama. You will keep our pipelines fast and our deployments predictable, and make it easy for engineers to ship safely without needing to understand how any of it works underneath.

Making things fail gracefully. You will work with the engineers building our services on how the platform behaves when something is slow, something is down, or a message cannot be processed. Retries, resource limits, queue handling and recovery, designed in rather than discovered during an incident.

Keeping the cost sensible. Our hosting cost grows as we add clients. You will work on right sizing, compute efficiency and understanding what each client environment actually costs us.

Incidents. You can volunteer to join a compensated, out of hours rota shared across the team, and take part in making sure the same incident does not happen twice.

• Engineers can answer normal operational questions without needing production access.
• We find out about problems from our monitoring rather than from a client.
• Deploying is routine enough that nobody plans their week around it.
• Adding a new client is routine and automated.
• Cost per client environment is understood and trending the right way.
• Nothing in the estate depends on one person knowing how it works, including you

• You have run production infrastructure on AWS for several years and know it properly: EKS, IAM and networking in particular, across multiple accounts and regions.
• You are strong with Kubernetes beyond the basics. Multi tenancy, Helm, scaling behaviour, and what actually happens when a cluster is under pressure.
• You write Terraform that other people can safely change, and you treat infrastructure code with the same standards as application code.
• You write code. Python, Go, or something comparable. Enough to automate properly, build the small services and controllers that fill the gaps, and read the applications you are supporting.
• You have built or run CI/CD properly, and have worked with Git Ops as a way of managing what is deployed.
• You have operated data heavy systems. Message queues, search clusters, pipelines that move a lot and cannot quietly lose anything.
• You…