Your transformation partner from inside the ministry. We bring senior AI know-how into your perimeter — and ship at operating speed.
One value, two delivery shapes: a scoped transformation partnership, or a standing senior engineering squad. Both put the expertise inside your perimeter.
A scoped engagement around one high-leverage AI workflow. You bring in senior partners with frontier-tech transformation experience — they sit inside the ministry and ship code with your team, not a deck at them. Sized in time and cost to the actual scope.
An ongoing senior engineering team alongside yours, month after month. Production discipline from Google-scale infrastructure paired with your engineers, transferring the practice as it ships. Paid against output, not against time on the clock.
The first wave of public-sector AI was procured through licenses and portals. Most of it never landed, because the hard part was never the model. The hard part is the work done inside the building, by people who have already done this work at frontier-tech scale. That is who we bring inside.
The model is well-tested in the private sector — see Tenex Labs, the exemplary framework for embedded AI transformation teams. Fieldspan Embedded brings the same shape to the institutions that run a country.
The people you meet in the briefing are the people on the ground in the engagement.
Built and ran the global installer network at SolarEdge across dozens of countries — a decade of making complex operational systems actually work outside the slide.
Also: Volunteers with the עיר-מדינה (Ir-Medina) civic initiative.
Spent 12 months on AI transformation at Google. The last 6 months advising on AI transformation externally.
The founders are on the engagement, not selling it. We ship working systems, not decks. And the IP belongs to you the moment it is written — source, weights, runbooks.
Two shapes. A transformation engagement is scoped jointly and sized in time and cost to the workflow. An engineering squad is a monthly subscription priced against output — you pay for features shipped into production, not for hours on a timesheet. No tokens, no per-seat licenses, no platform subscriptions.
Our people sit in your offices, on your devices, under your policy. Your data never leaves your perimeter, and it is not used to train anything for anyone else.
When it stops compounding for you. There is no fixed endpoint; most engagements roll from one workflow into the next, and each cycle ships faster than the last because your team is more capable and the operating practice is now resident inside the institution. If a cycle stops delivering, you stop the subscription — the code, weights, and runbooks are already in your repo from day one.