Companies of any size. The problem changes; the discipline does not.
A founder needs to know whether the thing can be built at all. A Fortune 100 needs forty teams to stop solving the same problem forty ways. Both are architecture questions, and the twenty years behind this practice covers everything between them.
The four
Four sizes. The problem changes with each one.
Open a stage for what it typically needs, how engagements run there, and where the depth comes from.
- Startup Pre-seed to Series A Feasibility, architecture and the technical side of a raise.
- SMB Roughly 10-200 staff Security leadership and a first audit, sized for the business.
- Mid-market Roughly 200-2,000 staff Multicloud architecture and a practice that outlasts me.
- Enterprise 2,000+ staff, including F100 Portfolio architecture, M&A integration and enablement at scale.
Most of this industry is specialized by size, and sensibly so.
Each part of it is built for a particular shape of problem. Offshore and nearshore firms sell staffed development capacity and are good at it. Large integrators are built for enterprise programs, and below that size the overhead is hard to justify. Boutiques know one industry properly, which is exactly why they stop at its edge.
On staffing specifically, I start from US-based teams, and often that is a condition of the funding rather than a preference: SBIR research has to be performed in the United States, and moving any part of it offshore takes written approval from the funding agency for a rare and specific reason. I have worked SBIR projects and so have the people I would bring in. Where no such rule applies, the rate card is still not the cost - time zones, lost context and review overhead show up later as rework. There are real reasons to go offshore anyway, and when one applies I have people I trust and I set the seams up carefully. What I will not do is pretend it is free.
I am not describing a gap in the market. I am describing three gaps in what I understood, and where I went to close them. Silicon, because I was writing software against hardware I could not see inside. Then years finding ways around other people's controls, because I could build and could not break. Then platforms at hyperscale, because I had never watched one decision land on thousands of estates. I worked each of those. I did not read about them.
Each of those left people behind, which is the whole reason the network reaches as far as it does. The silicon years are why I know process and RF engineers and the labs that test their parts. The research years are why I know assessors and penetration testers. The platform years are why I know SRE and DevSecOps teams. Founding companies is why I know VARs, resellers and the people who actually close a deal. I did not assemble the network. It accumulated, one job at a time - here is how far it runs.
Tell me what you are buildingWorked at all four
- Startup - the first architecture, before there is a team to inherit it
- SMB - standing security leadership and first audits
- Mid-market - multicloud programs and practice build-out
- Enterprise - portfolio architecture and post-merger integration
A single hosting portfolio once covered SMB, mid-market and enterprise simultaneously, across software, e-commerce, healthcare, legal, manufacturing, education and government.
What we build, and who else comes in
Most of this is work we do together. The network is what gets added when a job needs a specialist, not a hatch I push people through.
| What you are dealing with | What we build together | Who else comes in |
|---|---|---|
| You need a team that does not exist yet | The whole function, not just engineers: roles defined, candidates sifted, technical calls judged. Engineering, project and account management, research, compliance, technical writing, and sales if you are taking it to market. | Your own hires first, plus people from the network where you want them |
| You hold an SBIR award, or you are writing the proposal | The technical plan and the work itself, inside the rules: research performed in the United States, and the awardee doing the share the phase requires | US-based specialists who have delivered SBIR work, named in the proposal rather than found afterwards |
| You need volume, or cover across time zones | US-based teams first. Where the work genuinely has to be distributed, the structure that stops quality sliding: handoffs, acceptance criteria, and documentation that survives the gap between time zones | US teams first, and US only where the funding requires it. Nearshore and offshore where there is a reason - Poland, the UK, India, Mexico, Colombia, Brazil |
| Nobody owns delivery or the customer relationship | Project and account management built as a real function, with a reporting line a customer will accept | Operators from the network if you need them before you can hire |
| Custom silicon or RF inside the product | Architecture and advisory from inside the team. I have carried the flow to a fabricated netlist | A design house or test house when the fab relationship matters |
| A machine on a factory floor | Controls architecture, OT segmentation and safety boundaries | A controls integrator or machine builder for the install |
| Systems that move water or power | Generation, treatment, cogeneration and NERC CIP | Hydrogeologists, water scientists and power engineers |
| A security posture an acquirer will accept | Evidence design and the program behind it, built with your team | An assessor, who has to be independent of me |
| Forty architectures after three acquisitions | Portfolio alignment and integration, done through a real merger | Your architects, and extra hands if the timeline demands them |
| The documentation does not exist | Technical writing, and a structure that keeps it current after I leave | Writers who have worked under a regulator |
| You are not sure what you need yet | A call, no charge, and a straight answer about whether I am useful | Nobody yet. That is the point of the call |
When the answer is a bigger team than mine
A pure consumer web or mobile application, with no hardware, no regulator and no industrial context, needing eight developers by next month. That is a staffing model, and a US-based staffing firm is where I would start. I will say so in the first ten minutes rather than discover it in month three, and if you want it I will help you pick the team and build the handoffs that make it work.
Equally: a multi-year enterprise program needing hundreds of consultants on site. That is work for a large integrator or a specialist consultancy, and I would rather introduce you to one than pretend the shape fits.
Neither of those means I disappear. Picking the right delivery partner and setting up the seams is work, and it is work I am glad to do. Most of what arrives here is work we end up doing together anyway, with specialists brought in around it rather than instead of it.
- First step
- 30-minute call, no charge
- Then
- A budget scoping exercise, before anything is committed
- Startup and SMB
- Usually a focused piece of work rather than a standing arrangement
- Mid-market
- Quarterly engagement with knowledge transfer
- Enterprise
- Program engagement, alongside existing partners
- Introduction only
- No charge, no obligation