Staff Engineer — Performance, Reliability & AI Automation

Hey there Kubernetes people!

Every engineer here waits on CI several times a day. We want someone who finds that wait personally annoying.

Why this role exists

Hundreds of developers push to a large Ruby on Rails monorepo and every push lands on a runner fleet that the Developer Experience team builds runs and keeps fast.

That fleet is self-hosted Kubernetes on bare metal in European datacentres in the low hundreds of machines running ephemeral GitHub Actions runners that peak above a thousand concurrent pods. We own all of it: the hardware the clusters the runner images the caches and the delivery pipelines on top. We add nodes by hand rather than letting an autoscaler do it so capacity planning is genuinely part of the job.

The leverage is unusual. Shave a minute off the average build and every engineer in the company gets that minute back several times a day.

The mission

• Own the CI and CD Kubernetes clusters end to end: capacity reliability upgrades security and cost.
• Run the self-hosted GitHub Actions platform at scale with Actions Runner Controller. Runner scale sets ephemeral pods docker-in-docker and the runner images themselves.
• Keep the data services our tests depend on fast and healthy. MySQL Redis and ClickHouse come up per job alongside Rails application containers.
• Design and tune the caching that makes builds quick: node-local overlay image and artifact caches we build ourselves plus pull-through registry mirrors.
• Bring queue wait and build duration down with measurement behind it. Instrument the platform set targets and show the improvement.
• Manage it all as code. Terraform planned and applied from pull requests GitOps delivery with Flux and Argo CD Kustomize and Helm for manifests.
• Provision and operate bare metal. Linux networking across datacentre segments and troubleshooting that sometimes ends up at the disk or the kernel.
• Lead incident response for the platform run blameless postmortems and turn each one into an alert a guardrail or a runbook.
• Treat the platform as a product whose users are engineers. Talk to them watch where they get stuck and build paths that make the right thing easy.

Your day to day

• Rightsizing runner tiers against a week of CPU and memory data then opening the pull request that changes them.
• Tracking a flaky job from a red check down to the container runtime the kernel or a clock that drifted.
• Adding machines to the fleet: install enrol network verify document.
• Cutting p95 queue wait by working out which stage actually blocks.
• Upgrading a cluster or a controller without anyone noticing.
• Pairing with a product engineer on a workflow that has been slow for a reason nobody has looked at yet.

Your profile

• 5 years running production infrastructure with Kubernetes at the centre of it.
• Real depth in Kubernetes: scheduling requests and limits evictions node pressure DaemonSets controllers and operators. You have debugged a cluster that was lying to you.
• Strong Linux and container fundamentals. containerd or Docker internals cgroups namespaces storage drivers networking.
• Terraform and infrastructure as code with a feel for module design and state hygiene.
• CI/CD at the platform level: shared workflows reusable templates runner architecture build caching.
• Comfortable with GitHub Actions including self-hosted runners.
• You script to automate. Bash plus at least one of Python or Go.
• Observability as a working method. Metrics logs traces dashboards and alerts that people trust with OpenTelemetry and Prometheus style tooling.
• Operational maturity. On-call incident command postmortems and the discipline to chase causes.
• Capacity and cost awareness. You can size a fleet and explain the bill.
• Clear written English. You document what you build because the next person on call will not be you.

Bonus points

• Large Ruby on Rails test suites and what it takes to parallelise them honestly.
• Operating MySQL Redis or ClickHouse yourself.
• Bare metal: provisioning hardware failure and the different rhythm it has compared to cloud.
• Overlay and WireGuard style mesh networking VLANs and Kubernetes CNIs such as Cilium.
• Building internal developer tooling portals or platform APIs.
• Supply chain and build security. Runner isolation secret scoping image provenance.
• Working across more than one cloud. We use AWS Azure and Cloudflare alongside our own hardware.
• Contributions to open source infrastructure projects.

How we work

At Factorial we believe the best products are built when people come together in person to collaborate challenge ideas and move fast. Thats why our Engineering Team follows an office-first flexible approach.

We work on-site several days a week (80%) using that time to connect align and innovate as a team. However we also support remote work when it makes sense (20%) for deep focus or personal needs.

Perks and benefits

• High growth multicultural and friendly environment
• Continuous training and learning based on your needs
• Alan private health insurance
• Healthy life with Wellhub (gyms pools outdoor classes)
• Save expenses with Cobee
• Language classes with Preply
• Get the most out of your salary with Payflow
  And when at the office:
• Breakfast in the office and organic fruit
• Nora and Apeteat discounts
• Pet friendly

About Factorial

Factorial is an innovative Business Management Software solution designed to streamline company processes for small and medium-sized enterprises. Founded in 2016 our mission is to help companies automate workflows centralize people data and make better business decisions. With customers across over 60 countries weve built a diverse team thats driving change in the tech space.

Our values

• We own it: We take responsibility for every project.
• We learn and teach: We share knowledge every day.
• We partner: Every decision is a team decision.
• We grow fast: We act fast and learn from our mistakes.

Diversity

Diversity is part of our culture. We have more than 43 nationalities and provide an inclusive environment for all. Please feel free to apply however suits you best (blind resume identity pronouns cover letter etc.). We do not discriminate; we encourage everyone to join us!