Cloud infrastructure & reliability/Est. 2019
We're a two-person consultancy for teams of 8 to 60 running on AWS, GCP or Azure. We fix the parts that wake people up at night, and we stop the bill from creeping up six percent a month while nobody is looking.
Fixed-fee engagements · senior engineers only · three clients at a time
No reseller deals, no referral fees, no bench of juniors learning on your account.
Everything else, we'll tell you to keep in-house. A two-person firm that pretends to do everything is a two-person firm that does nothing well.
Two weeks inside your infrastructure, read-only. You get a written review you can hand to your board, a risk register ordered by what would actually hurt, and a 90-day plan your own engineers can execute without us.
2 weeks · fixed fee · from $4,800Not "turn things off". We map spend to services and teams, kill the orphaned resources, right-size the rest, then put budgets and anomaly alerts in place so the number stays down after we leave.
3–5 weeks · fixed fee · from $9,500SLOs that mean something, an on-call rotation a human can survive, runbooks written by someone who has been paged at 3am, and a game day to prove it all works before your customers find out it doesn't.
4 weeks · fixed fee · from $14,000Cloud-to-cloud or data-centre-to-cloud, with a rollback plan that has been rehearsed. If the honest answer is "don't migrate yet", that's what you'll hear — in writing, with the numbers behind it.
6–10 weeks · scoped · from $24,000Names withheld under contract. Numbers aren't — they're the only part worth reading.
All three from engagements in the last 18 months, quoted with permission.
"They didn't sell us a platform. They fixed what we already had, wrote down why, and left. Our bill is 41% lower and nobody has been paged at night in a month."
"Our first real outage happened on a Friday afternoon. The runbook Nimbus wrote for us is the reason it lasted 40 minutes instead of a weekend."
"I've worked with three consultancies. These two are the first who told us not to do the project. We hired them again the following quarter."
How we work
We're two people. There is no account manager between you and the engineer doing the work, because there is no account manager. Everything we recommend, we have run ourselves in production — including the parts that failed.
We write things down. Every engagement ends with documents your team owns: architecture decisions, runbooks, a cost model, and an honest list of what we'd do next if we had another month. If you never hire us again, those documents should still be useful in a year.
Bring your bill and one thing that keeps breaking. You'll leave with two fixes you can make this week — whether or not you hire us.