The engineer who can talk to the customer, come back, and build it.

Software is getting cheap. Deployment is not.

For most of the history of software, the expensive part was writing it. Companies organized themselves around that scarcity: product to decide, engineering to build, sales to explain, support to absorb the difference. Every layer existed because the skills on either side of it didn't transfer.

AI collapses the cost of the middle. Code is no longer the bottleneck. What remains scarce is the person who can sit in a real operation, understand what is actually broken, decide what should be automated and what shouldn't, build it against the systems that are really there, prove it works, and say what it was worth in terms the business understands.

That person has a name now. The Forward Deployed Engineer started as a job title at AI companies. It is becoming the shape of engineering itself.

Where the work goes next.

The demand is arriving in waves, and each one is larger than the last.

Nobody is trained for this. Everybody is interviewed for it.

The skills the role demands — discovery, scoping, evaluation, integration, communication — are not taught in computer science programs and not taught on most engineering teams. Hiring loops test for them anyway. The result is a market with far more open seats than people who can credibly fill them, and thousands of capable engineers who could, but have never done a deployment end to end and can't prove it.

The gap isn't intelligence or experience. It is reps: one real problem, found and scoped in a real business, built and shipped into real systems, measured, and written up. Most engineers never get that first rep inside their current job. The program exists to make sure they get it.

Taught the only way it can be: by doing it.

There is no lecture that produces a deployment engineer. FDE Dispatch programs are built around a live deployment for a real organization, coached one-to-one by someone who does the work for a living, with the interview loop practiced alongside so the proof is legible to the people who hire.

Everything a participant leaves with is evidence, not credentials: a system running in production, a case study written the way deployment teams write them, and recorded interview performance that shows the change. The program is small on purpose. Rigor doesn't scale by adding seats; it scales by making every seat count.

A dispatch first. Then a guild.

The name is deliberate. A dispatch is a report from the field, and the field is where this discipline is being written — one deployment at a time, by practitioners who mostly work alone. The program is the first chapter: turn capable engineers into deployment engineers, and document what works.

What follows is the network: the patterns, the evaluation methods, the reference architectures, and the people who carry them from one organization to the next. The engineers who deploy AI into the world should not have to learn it alone. That is the mission.