About the role
CloudPeak's control plane is the part of the product customers never see and cannot operate without. It provisions landing zones, enforces guardrails, reconciles drift, and produces the evidence pack an auditor asks for six months after a migration nobody remembers. It is around 180,000 lines of Go, deployed into forty-two regulated environments, and it is now the busiest codebase in the company.
We are adding a senior engineer because a reconciliation loop that was comfortable at ten tenants is creaking at forty. You would own that path end to end — the API, the workers, and the PostgreSQL schema underneath — working alongside four engineers, an SRE embedded from the reliability team, and a product manager who spent eight years running infrastructure inside a local insurer.
Expect design documents that get argued over, code review that catches things, and change windows that start at 2am roughly four times a year. Expect also that nobody will ask you to sit through a status meeting you could have read in four minutes.
What you'll do
- Design, build and operate Go services in the control plane, from HTTP and gRPC edges through to the job workers that do the real work.
- Own the reconciliation pipeline that keeps forty-plus tenant environments matching their declared state, including drift detection and safe remediation.
- Model and evolve PostgreSQL schemas, including zero-downtime migrations against databases that customers will not let us take offline.
- Bring the provisioning API's p99 back under two seconds as tenant count grows, with the traces and benchmarks to prove it.
- Write the immutable audit trails and evidence exports that HKMA- and HKIA-supervised customers depend on at review time.
- Review designs and code across the team, and mentor two mid-level engineers who are ready for more scope.
- Take part in the paid on-call rotation for the services you own — one week in six, with a genuine handover.
What we're looking for
- At least five years building backend services in production, two or more of them in Go.
- Real PostgreSQL depth: query plans, locking behaviour, and migrations you have run against live data.
- Kubernetes experience as both a user and a platform builder — you have written controllers or operators, not just YAML.
- You have owned a service in production, debugged it at 3am, and written the postmortem afterwards.
- Business-level English; the team designs, reviews and documents in English.
Nice to have
- Exposure to regulated workloads — HKMA cloud guidance, HKIA supervisory expectations, or PCI DSS.
- Published Terraform providers or Kubernetes operators.
- CKA or CKAD — and if you do not have it, we pay for it in your first year.
What you get
- Certification exams and training fully funded (AWS, Azure, CKA)
- Medical and dental cover extending to spouse and children
- MPF employer contribution topped up to 9%
- 20 days annual leave, rising to 25
- Hybrid — two days a week on site, shuttle bus from University station
- Paid on-call allowance on top of base
Skills & keywords
Marcus Tang
Head of Platform Engineering · reviews applications personally
Listing ID JOB-TECH-001 · Closes 4 Sept 2026 · HKjobs never asks candidates to pay a fee. Report this listing