Process
Measure. Design. Build. Leave.
Four stages, in that order. The order is the method: each stage's output is the next one's input, and skipping the first is how teams spend a quarter optimising something that was never the bottleneck.
- 01
Diagnostic
1–2 weeks
We measure before we recommend. Your workload, your traffic, your costs and your constraints, turned into a picture of where the time and money actually go — which is regularly not where the team assumed.
Leaves you with
- A measured baseline: latency, throughput, cost per request
- The bottleneck, named and evidenced
- Options with trade-offs, including the option to do nothing
- 02
Architecture
1–3 weeks
The design, written down and argued through with your engineers. Decisions get recorded with their reasoning, so the team inheriting this can tell a deliberate choice from an accident.
Leaves you with
- A reference architecture your team has reviewed
- Decision records covering what was chosen and what was rejected
- A sequenced plan with the risks called out
- 03
Build
4–12 weeks
Hands on keyboards, in your repository, against your review process. We work alongside your engineers rather than beside them — if the knowledge leaves when we do, the engagement failed.
Leaves you with
- Working systems in your infrastructure, in your IaC
- Evaluation and load tests running in your CI
- Observability wired into the stack you already use
- 04
Handover
1–2 weeks
The part that gets skipped, and the reason systems rot six months later. Your team operates it while we are still available to answer for it.
Leaves you with
- Runbooks for the failures we can anticipate
- Working sessions with the engineers taking it on
- A named list of what we would do next, and what we would not
Engagements
Three ways to work with us.
Diagnostic
1–2 weeks
You suspect something is wrong but cannot name it.
A fixed-scope measurement of your workload, ending in a written picture of where time and money go, and what we would do about it. It stands alone — plenty of teams take the report and do the work themselves.
Project
6–14 weeks
You know what needs building and want it built properly.
Diagnostic, architecture, build and handover against an agreed outcome. Our engineers work in your repository alongside yours, under your review process.
Retained advisory
Ongoing, monthly
Your team is capable and wants a second pair of eyes.
Regular review, architecture support and an escalation path when something breaks. Deliberately part-time — if you need us every day, one of the other two is the honest recommendation.
Pricing is fixed for the diagnostic and quoted per project after it. We do not quote a build before we have measured the thing we are quoting for.
How we operate
Four commitments.
Measure first
No recommendation before a baseline. Teams routinely optimise the wrong thing for a quarter because nobody put a number on it in week one.
Your repo, your cloud
Work lands in your infrastructure, your IaC and your review process from day one. Nothing of consequence should live on a consultant's laptop.
The boring answer, when it's right
Frequently the correct advice is a cache, a shorter prompt, or staying on the API you already pay for. We would rather be useful than interesting.
Leave the team stronger
Success is your engineers operating this confidently without us. We are a temporary function by design.
Questions
Asked often.
Do you take equity or work for free up front?
No. Engagements are paid and scoped, starting with the diagnostic. It keeps the advice independent — including the advice that you do not need us.
Will you write code, or just advise?
Both, and the build is the bulk of most engagements. Work lands in your repository under your review process. Advice that never survives contact with a codebase is not worth paying for.
What if the diagnostic says we don't have a problem?
Then it says that, and the engagement ends there. It happens, and it is a good outcome — you have a measured baseline and a reason to stop worrying.
Can you work with our existing vendors?
Yes. We have no reseller relationships and take no referral fees, so there is nothing for us in recommending one platform over another. That independence is the point.
How do you handle our code and data?
Under your agreements, in your environment, with the minimum access needed to do the work. For regulated engagements we work inside your boundary from day one rather than asking for an export.
Tell us what's breaking.
Bring a workload, a latency target, or a bill you can't explain. First conversation is a technical one — no discovery deck.