Three ways to start
Most clients begin with a fixed-scope pilot. The other two exist because not every project starts from nothing, and not every roadmap fits in one engagement.
Architecture review
For teams who have a build underway and want a second opinion before it scales.
- Code and architecture walkthrough
- Load and failure-mode analysis
- Security review of device identity and OTA
- Written findings, ranked by risk
Build a pilot
The usual starting point. A working system on a real site, priced up front.
- Firmware, backend and console
- 20–50 device pilot fleet
- Cloud environment you own
- Weekly demo, not a status report
- Source, docs and runbooks on delivery
Dedicated pod
For a long roadmap. A standing team that works as part of yours.
- Firmware, Node.js and React engineers
- Your rituals, your board, your repo
- Named engineers, no rotation
- Scale up or down with a month's notice
What actually differs between them
| Architecture review | Build a pilot | Dedicated pod | |
|---|---|---|---|
| Typical length | 2 weeks | 10–14 weeks | Rolling, 3-month minimum |
| Pricing | Fixed fee | Fixed scope, fixed price | Monthly rate per engineer |
| Team size | 2 engineers | 4–6 engineers | 3–6, adjustable |
| Firmware included | Reviewed only | Yes | Yes |
| Hardware in the field | Not included | 20–50 units | Depends on roadmap |
| You own the code | Included | Included | Included |
| Runbooks and docs | Findings report | Included | Included |
| Weekly demo | Final readout | Included | Included |
| Can convert to the next stage | Into a pilot | Into a pod | — |
When each one is the wrong choice
The fastest way to pick is usually to rule two out.
Architecture review
Two weeks, two engineers, one written report. We read the code, run the load profile you expect at scale, and try to break the device identity and update path. You get findings ranked by what they will cost you if left alone, with the cheap fixes separated from the structural ones.
You have something running, growth is coming, and nobody internally has the distance to assess it.
You have nothing built yet. Start with discovery instead — there is nothing to review.
Build a pilot
The whole path from sensor to screen, on a real site, with a fleet small enough to fix. Fixed scope and fixed price, agreed after a discovery phase so the number means something. Most clients start here, and most production builds are a continuation of the pilot rather than a rewrite of it.
You know what you want to measure and need to prove the whole chain works before committing to volume.
The requirement is still moving weekly. Do discovery first; fixed scope needs a fixed question.
Dedicated pod
A standing team inside your process: your standups, your board, your repository. Named engineers who stay on the account rather than a rotating bench. Typically three to six people, adjustable monthly, with a three-month minimum so the team is worth onboarding.
You have a roadmap longer than one project and you would rather grow capacity than hire for it immediately.
You need one specific thing built and finished. A fixed-scope pilot will cost you less.
Size the work before you talk to anyone
Pick what needs building and the estimator returns a duration range, a team size and the engagement model that usually fits. It is deliberately rough.
Rough shape of the engagement
- Estimated duration
- 15–20 weeks
- Team size
- 4–5 engineers
- Effort before parallelism
- 16 eng-weeks
- Suggested model
- Build a pilot
This is a budgeting range, not a quote. It exists so you can take a defensible number into a planning conversation. Discovery turns it into a fixed price we will hold to.
Send this scope to usA number produced before anyone has seen your constraints has to be padded to be safe, and padding is how projects get priced out of existence. Discovery is two weeks and priced separately, and the fixed price that comes out of it is one we hold to.
Bring the estimate and we will argue with it
The most useful first call is one where you already have a rough shape in mind. We will tell you which parts we would cut.