About The Team
Sea Cloud gives businesses and developers the infrastructure to build, train and run AI at scale: GPU compute, storage, networking, managed Kubernetes and model APIs, delivered as a public cloud.
The Compute and Storage SRE team delivers and provisions the platform's foundation products and keeps them available, fast and durable: virtual machines, managed Kubernetes, block storage, object storage, and the traffic gateways that put them in front of customers. We combine software and systems engineering: we build the automation and observability that run these services, define and defend service-level objectives, write the standard operating procedures (SOPs) the platform runs on, and lead the response when things go wrong. We also prove resilience rather than assume it, through chaos engineering and platform-level disaster-recovery fire drills, and we harden the platform's security and strengthen its compliance controls.
Join us to help build and run a cloud platform, where your work shapes how the service is run for years to come.
Job Description
- Own the reliability of one or more compute and storage products, such as virtual machines, managed Kubernetes, block storage, object storage or traffic gateways. You set their SLOs, agree error budgets with the product teams, and report against them.
- Own the delivery and provisioning of infrastructure products: design repeatable, automated pipelines that bring up new capacity, regions and availability zones (hosts, hypervisors, Kubernetes and storage clusters, traffic gateways), with acceptance tests before anything goes into service.
- Design and implement systems that continually improve the reliability, efficiency and security of the compute and storage platform, from architecture through to production code.
- Review new designs and launches for failure modes, data durability, observability and recovery before they reach customers.
- Design and build the reliability tooling the team runs on: automated detection and remediation, safe rollouts with canaries and automatic rollback, and capacity forecasting for compute and storage.
- Own the SOP library for your products: make every SOP current, tested in drills, and automated where it can be.
- Lead the chaos engineering programme: choose the failure modes that matter (host, rack, disk, network partition, gateway, control plane), design experiments with a bounded blast radius, and turn findings into fixes.
- Plan and run platform-level disaster-recovery fire drills, up to the loss of an availability zone. Measure recovery against RTO and RPO targets and keep the evidence auditors will ask for.
- Drive security hardening and compliance for your products: define hardening baselines for hosts, hypervisors, Kubernetes and storage (for example CIS benchmarks), run vulnerability and patch management, enforce least-privilege access and encryption, and turn compliance controls into automated checks and audit evidence.
- Lead major incidents as incident commander, run blameless post-incident reviews, and drive the fixes to closure.
- Mentor graduate engineers and shape the on-call rotation.
Requirements
What you need
- Minimum 2 years in SRE, production engineering, infrastructure software engineering, or a related field, on large distributed systems.
- Strong coding in Go or Python. You have shipped production tools and services, not only scripts.
- Production experience operating at least one of: a virtualization platform, Kubernetes at scale, or a distributed storage system.
- Deep Linux skills: kernel, file system and block I/O behaviour, performance analysis, and debugging under pressure.
- Infrastructure as code (Terraform or similar) and building an observability stack, not only using one.
- A track record of leading incidents and of running failure tests or DR exercises in production.
- Working knowledge of datacenter networking: L3 fabrics, BGP, load balancing, DNS.
- Logical reasoning and critical thinking: you question assumptions, weigh the evidence, and can explain your conclusions.
- Strong ownership: you see a problem through to the fix, not just to the handoff.
- Eagerness to learn: you pick up new systems quickly and keep growing your skills.
Nice to have
- Virtualization internals: KVM, QEMU, libvirt, live migration, or a cloud stack such as OpenStack.
- Distributed storage internals: Ceph (RBD, RGW), an S3-compatible object store, erasure coding, replication and rebuild behaviour.
- Gateways and load balancers in front of cloud APIs: S3 gateways, API gateways, L4/L7 load balancers, rate limiting.
- Chaos engineering tools such as Chaos Mesh, LitmusChaos or Gremlin, or a fault-injection framework you built.
- Security hardening in practice: CIS benchmarks, OS and Kubernetes hardening, secrets management, policy-as-code (OPA, Kyverno).
- Running a public cloud service with external customers, SLAs and compliance audits (ISO 27001, SOC 2, MTCS).
- GPU servers or AI workloads on the compute and storage you run.