03/About · The practice
We held the chair. We never stopped shipping.
Most people advising on AI-era engineering have done one of two things: run teams, or build with the tools. The whole point of this practice is that we do both — and the advice is better for it.
We’ve run engineering as CTO and VP — owned the headcount, the budget, the hiring bar, and the 2am calls. We know what it feels like to be accountable for output you can no longer cleanly measure, and to watch good people burn out carrying work the org chart never accounted for.
And we never moved into the management track and stopped building. We still ship production software today — native mobile, embedded, the AI/ML backend, and on-device edge AI with the UI/UX wrapped around it — on AI-augmented workflows we run every day. So when we tell you where the bottleneck moves or why a metric stops meaning what it meant, it’s not from a whitepaper. It’s from this week.
There’s a sharper edge to that. Used well, AI turns an ordinary developer into something close to 10x — we’re the proof, and we’re also the cautionary tale: the amplified developer who realised he could ship alone, and did. So we know the part most people miss. Scaling your developers with AI is the easy half. Keeping them — managed well, still growing, and loyal to the team rather than tempted to go it alone — is the half that decides whether the productivity is yours, or walks out the door.
That dual vantage point is the practice. It’s also why the advice is vendor-neutral — no AI tool to resell, no kickback, no allegiance — and why it’s anchored to before-and-after numbers rather than vibes. Every engagement is built to be defensible against an outcome, starting with our own.
—/Expertise
The stack most AI consultancies can't touch.
The leadership advice is credible because the engineering is real. We work across the hardest parts of the field — and lead teams that do the same:
- Native mobile
- iOS and Android shipped to production — not web wrapped in a shell.
- Embedded & edge
- Software on constrained devices, where every millisecond and milliwatt is accounted for.
- Backend AI/ML
- The training, serving, and data plumbing behind models that actually run in production.
- On-device / edge AI
- The rare part: AI and ML running on the device itself — private, offline-capable, low-latency — not round-tripping to someone else's cloud.
- UI/UX
- Making all of it usable. On-device intelligence is worth nothing if the product around it isn't.
On-device and edge AI is the flagship — and one of the scarcest specialisms in the field. It’s also the clearest reason the advice carries weight on mobile, native, and embedded teams, where generic AI consultancies have nothing to say.
—/Our proving ground
The proof is in our own builds.
Before we ask you to trust a number, here’s ours. Notturno and Dekker are real production applications, shipped on the same remote agentic workflow we bring to client teams — an iPad driving a desktop build environment. They’re the laboratory: we feel every shift these tools cause, in production, before we bring it to your team.
- Real production apps
- Shipped end to end, native mobile and backend included — the work most AI consultancies quietly wave away.
- The same workflow we teach
- We don't recommend a way of working we don't run ourselves, daily.
- Where the bottleneck moved
- Tracked once writing the code got cheap, so we can show you where the hours go next.
The quantified before-and-after — cycle time, what one person can now ship, and exactly where the hours moved — is the anchor we’ll walk you through in a conversation, with the real production data behind it.
—/Working together
Instruments, not slideware.
You leave with something that works.
A live picture of team health, a manager’s handbook, a real interview loop, a governed workflow. Editable working documents you can run live with your team — not locked PDFs that die in a drive.
Direct, practitioner, no hype.
No “AI-powered” gloss, no transformation theatre. We’ll tell you what’s working, what isn’t, and what to do on Monday.
Your call, always.
We bring the judgement and the instruments; the decisions stay yours. The work is designed to make you a better leader of this — not dependent on us.
Start here
Start with where you actually stand.
Four minutes, four pillars, a maturity read and your most urgent next move. Then we talk.