- consulting
- engineering studio
- business
Why the engineering studio model works
Notes on running a lean technology consultancy: a small senior core, a vetted network of specialists, and an explicit choice not to scale into an agency.
There's a fork every technical consultancy hits early. You can scale into an agency — layers of project managers, junior engineers, account leads, all priced to make the org chart work — or you can stay a studio.
We chose studio, and we think most technical clients are better served by that choice. Here's why.
The unit of work is a small team, not a department
A two-to-four-engineer team that has shipped together can hold an entire system in its head. Architecture, code, deployment, monitoring, handover. Once you cross into double digits, the team stops being one brain and becomes a coordination problem.
Most of the work clients actually need — production ML systems, custom software, embedded integrations — fits in the first shape, not the second.
A network beats a payroll
When a project genuinely needs a specialist — a security auditor, a hardware engineer for a specific platform, a niche ML researcher — we bring one in from our network. They're typically people we've worked with directly before. They show up for the part of the project that needs them and leave when it's done.
The payroll alternative is to keep one of each on staff, charge for their downtime through bench rates, and then assign them to projects they're not the best fit for because they're available. We've seen that movie. We don't want to be in it.
Senior-only is a feature
Every engineer on a Voight Labs project is someone the client would have hired if they were available. No off-shore junior tier. No tickets written by a project manager who doesn't understand the system. The people on the call are the people writing the code.
This sets an upper bound on how many engagements we can run at once, and that's the trade we want. We'd rather decline a project than run it poorly.
Ship-oriented by default
Studios that bill for deliverables drift toward producing deliverables. We bill against engagements with a clear shape and a production outcome. If a system isn't running in production, it isn't done. We measure ourselves against that line.
Honest about scope
Some problems are the wrong shape for outside engineering. They need an internal team because the surrounding system is changing too fast, or because the institutional knowledge required can't be transferred in a quarter. When that's true, we say so on the discovery call.
It's a strange thing to say out loud — that we'll pass on work we could technically do — but it's the studio model's most underrated feature.
If this is the kind of engineering partner you're looking for, get in touch.