Small team.
Lots to do.
Metalhost is founder-led and deliberately small. We're not actively hiring as of today, but that changes the moment the platform crosses the next utilization line. If the work below sounds like yours, drop us a note — careers@metalhost.net.
Nothing posted right now.
We'd rather have no roles posted than fake openings to seem bigger. When we start a search, it'll show up here with the actual scope, comp range, and how to apply. Until then, the best move is to send something specific.
We try to reply to every signal-rich note. We do not reply to recruiters or mass-mailed CVs. No offense intended.
What we'll be hiring for post-launch.
Systems / SRE engineer
Linux, networking, storage, KubeVirt, OVN-Kubernetes, Tinkerbell. You're as comfortable in a switch CLI as in Go. On-call rotation when fleet size justifies it.
Platform / Go engineer
Connect-RPC + Postgres + Temporal day in, day out. Strong product instincts; you've shipped customer-facing APIs that other people read.
Frontend engineer
React 19 + TanStack Router. Comfortable building dashboards on top of generated Protobuf clients. Cares about live pricing, real loading states, and keyboard navigation.
Datacenter / hardware operator
Hands on Dell iDRAC, Tinkerbell PXE flows, Ceph storage layout, switch configuration (OS10 / SONiC). Local to a DC we operate in, or willing to travel.
Support / DX engineer
Email-first support that customers actually thank us for. You can debug a kernel panic over a screen-share, write the postmortem, and turn it into docs.
Operations / billing engineer
You like spreadsheets and double-entry ledgers as much as you like Go. Stripe + Coinbase Commerce + Postgres ledger reconciliation, tax handling, invoice automation.
Some of this matters more than a CV.
Owners, not ticket-closers.
We're small enough that there is no "someone else." The person who notices the problem fixes it, writes the post-mortem, and updates the docs. If that's exhausting, we're not the fit.
Write things down.
ADRs, runbooks, customer-facing docs. If it isn't written, it didn't happen. We over-document on purpose so future-us can rebuild from the docs alone.
Plain language.
Talk to customers and to each other like adults. No marketing jargon, no compliance theater, no fake confidence. "I don't know yet" is a complete answer.
Care about the bill.
We sell infrastructure on price + honesty. Decisions that quietly cost customers more without making the product clearly better are out of scope.
Read the source.
Stack overflow is fine; the underlying upstream source is the answer. We default to reading what the project actually does, not what a vendor blog says.
Boring tech, novel product.
Postgres + Go + Kubernetes + KubeVirt are deliberately unglamorous. The novelty is in the product, the pricing, and the operational discipline — not the stack.