Forward Deployed Engineering

Forward deployed engineering places software engineers inside a customer’s team for a defined period to build and change systems in the customer’s own environment, and holds them accountable for what is running when they leave. The term comes from military vocabulary and was in use at Palantir by 2009. In 2026 AWS, OpenAI, Anthropic, Microsoft and Google all stood up forward deployed teams. Our explainer What is a forward deployed engineer? covers the role, its history and how it differs from a consultant.

Consulting produces recommendations for your team to implement. A Mission implements. Our engineers make the changes in your repositories and environments, with your engineers, and the Mission is judged on what is running in week four rather than on a document.

Augmentation supplies engineers to work your backlog your way, for as long as you need them. A Mission brings its own method, a fixed length, one agreed objective and accountability for the outcome. It ends on purpose, with your team able to continue.

Managed Services runs your platform for you, continuously. A Mission improves a platform your team runs and then hands it back. Many Missions end with a backlog that becomes Managed Services work, and that path is open, but it is a separate decision.

One primary engineering team, one technical owner, one reasonably consistent architecture, one operating model and one clearly defined domain. A Kubernetes platform, a business unit’s cloud platform, an internal developer platform, a CI/CD or observability platform, a data or AI platform, a landing zone. Several environments count as one platform when the same team operates them on the same architecture.

No. A Mission works on the platform your team owns, on AWS, Microsoft Azure or Google Cloud, on Kubernetes wherever it is hosted, on data and AI platforms, and on your own infrastructure. Dedicatted is an AWS Premier Tier Services Partner and also works with Microsoft Azure and Google Cloud. The method does not change with the platform, only the engineering does.

A defined amount, which is the point. The commitment is three to five priority engineering items on one platform, with more if capacity allows, and every change measured. Reducing Kubernetes cost, tightening a security posture, cutting alert noise or taking an AI pilot to production are all four-week shapes. Optimising an entire cloud estate is not, and we say so at scoping.

You do, from the first commit. Everything is written into your repositories, under your review and your controls. There is no handover at the end because nothing was held back.

Yes, when the Mission needs it. The two engineers work inside your team’s working day, on site or remote, in your standups, your channels and your tooling. Specialists join remotely for the part of the Mission that needs them.

The Mission ends with a prioritised backlog you own. It can become the next Mission, a scoped project, a retainer, or Managed Services. Or nothing further, with your team continuing on its own. All four are fine outcomes.

One fixed price for the two-person, four-week Mission, agreed before it starts. Specialists are priced into the scope when the Mission needs them. There is no hourly billing and no change-order cycle.