Site Reliability Engineer

Who we are

Athos Commerce enables ecommerce and digital marketing teams greater control over how their products are found, merchandised, and sold - on-site and across all digital marketing channels. Our tools span AI-powered search, personalized recommendations, smart merchandising, optimised product data syndication, A/B Testing and marketplace/social commerce performance.

We help brands grow with sharper visibility, higher conversion, and less friction. Backed by PSG Equity and supported by a global team across The UK, North America, and APAC, we are building practical, powerful tools for the next era of commerce. Athos Commerce unites shopper intent and product visibility across every touchpoint.

Who You Are

You are a hands-on reliability engineer who has carried a pager and learned from it. You have been the person who worked out what was actually broken at an awkward hour, and you have built the thing that meant it did not happen the same way twice. You are comfortable with distributed, asynchronous systems where failure is rarely obvious and partial success is the normal failure mode.

You are pragmatic about monitoring. You know that the right number of alerts is usually lower than the current number, that a success message is a claim rather than evidence, and that silence only means something if you have made it mean something. You inherit someone else's work and improve it without disparaging it. You take responsibility for what you build: you follow a change into production, check it did what you expected, and say so first when it did not. You move between a query, a pipeline and a pull request in the same week. You are motivated by systems that quietly keep working, and by the discipline it takes to get there.

What You'll Do

As our Site Reliability Engineer, you will be the first dedicated SRE in a newly formed Platform and Site Reliability Engineering function, reporting to its Lead.

We run more than 100 services in production, moving very large volumes of retail data through an asynchronous pipeline: uploads land on a queue, workers transform in batches, scheduled jobs build outbound feeds, and syndication services push to partner APIs that change without warning.

We already have a substantial observability estate — Azure Monitor, Log Analytics, Application Insights, Grafana, OpsGenie paging, external uptime checks and runbooks — built by engineers who had services to ship and did the right thing on the way past.

The Lead owns it and has assessed where it holds up and where it does not; what the function lacks is the hands to act on that. Your work is to make it trustworthy, and to bring your own read of what should change rather than waiting to be told.

The function covers reliability, the engineering platform beneath it, and the automated testing that guards the way into production, so alongside production signal you would be writing test automation and building the gates and pipelines that decide whether a release is safe to ship.

A growing part of that is how we use AI tooling inside engineering, which you would help build into the way the team works.

Occasionally you would pick up feature or bug work alongside a stretched delivery team, which keeps you close to the code you are keeping reliable.

This is a hands-on role rather than an architecture or management position: you will be writing queries, alert definitions, test code, pipeline configuration and application code every week.

How You Will Succeed

Improve Production Signal: Work out which alerts earn their place and which do not, recommend the changes, and build what is agreed. The aim is fewer pages, each of which someone acts on rather than acknowledges.

Make Silence Trustworthy: Build the mechanisms that separate "nothing is wrong" from "nothing is reporting" — heartbeat alerting, detection of disabled or non-firing rules, and synthetic checks that prove the alerting path still works end to end.

Support and Help Define Service Level Objectives: Work with the Lead and the owning teams to define targets, deriving candidates from real data for the feed paths customers notice first. Then build what makes them mean something: the instrumentation, the reporting that keeps the number in front of people, and the error-budget alerting that makes a breach something we act on at the time. A target nobody measures is a preference.

Support Incident Response: Take a vague report and narrow it down across API, queue, database and partner integration. Write up what happened and what should change, and return later to confirm the change held.

Bring Alerting Under Engineering Discipline: Move alert definitions and thresholds into version control and under review, alongside the services they watch, so that a disabled rule or an altered threshold has an author and a history.

Make a Green Build Mean Something: Write and maintain the automated tests that stand between a change and production, and work out where our current suites would pass while the system was broken. Build contract tests that catch a partner changing their API before a customer reports it.

Improve How We Ship: Support the Lead in building the paved road our engineers ship on — the frameworks, templates, pipelines and shared tooling that make the quick way and the safe way the same way. Much of this runs through our use of AI tooling: building the agent skills that help engineers produce and maintain test coverage, and judging where that tooling has earned trust and where it still needs a human gate.

Reduce Toil: Find the repeated manual work our reliability currently depends on — the checks someone runs by hand, the reports someone assembles every week — and automate it away. This one is yours to drive.

Build for Other People: Produce dashboards engineers actually open, runbooks responders actually follow, and queries that answer the question rather than needing translation. Nothing important should depend on you being awake.

Make the Case for Change: Identify where our reliability practice lags behind the scale we operate at, and make the case with the evidence behind it. Several of the improvements will need buy-in from teams you do not sit in.

What We're Looking For

• 5+ years of experience in site reliability, platform, infrastructure or DevOps engineering, with genuine ownership of production systems.
• Experience on a production on-call rota, responding to live incidents and living with the consequences of your own alerting.
• A track record of tuning alerting rather than only creating it, including removing, merging or rewriting alerts that were not earning their place.
• Strong query skills you use unaided: telemetry in KQL, PromQL, SPL, Datadog query language or similar, and SQL against a real schema to confirm what an application actually stored.
• Hands-on depth in a major cloud platform. Microsoft Azure strongly preferred; equivalent AWS or Google Cloud experience transfers.
• Experience owning CI/CD pipeline configuration, and the ability to read a failing run and tell whether the cause is the code, the test, the infrastructure or the environment.
• Professional C# development experience, writing production code to a standard that passes peer review. This one is not negotiable: our platform is .NET and you will be working in the application, its test automation and the tooling around both.
• Proficiency in a scripting language you write unaided, such as PowerShell, Python or Bash.
• Practical experience of distributed, asynchronous systems where partial failure is common — queues, batch processing, scheduled jobs or third-party integrations.
• Experience writing and maintaining automated tests yourself, whether integration, contract or end-to-end, with a view on what is worth gating a release on. A passing suite and a quiet alert are claims rather than evidence.
• A record of reducing toil through automation, with a before and after you can describe. Tooling other engineers adopted and kept using counts for more than a script only you ran.
• Excellent written communication, including postmortems, proposals and runbooks that colleagues act on.
• Serious, current use of AI tooling in your own engineering work, and an interest in building it into how a team works rather than only using it yourself. A considered view on where it helps and where it misleads matters more than enthusiasm.
• Service level objectives and error budgets in practice, alerting defined as code, Service Fabric or Kubernetes, contract testing against third-party APIs, or high-volume data pipelines is advantageous.
• The right to work in the UK — we are unable to offer visa sponsorship for this role — and located within approximately two hours' travel of central London, able to attend our Victoria office around once a month.

Benefits

• Competitive compensation
• 25 days' holiday per year, increasing by one day for each additional year of service, to a maximum of 28 days
• Company pension scheme
• Vitality Healthcare/Half Price Gym membership/Free Coffee and more…
• Cycle to work program
• Flexible working between home and office

Athos Commerce is an equal-opportunity employer and proud to foster a workplace free from discrimination and harassment. We strongly believe that diverse backgrounds, experiences, and perspectives are essential in cultivating an inclusive culture and building an innovative, successful organisation. All qualified applicants are considered for employment, without regard to age, race, colour, religion, sex, national origin, disability, veteran status, sexual orientation, gender identity, or any other protected status.