Encord's head of FDE on building the function: engineers first, pods with deployment strategists, and a 20% goodwill rule on scope
In a Nov 9 interview, Encord's Head of Forward Deployed Engineering Justin Sharps explains how he turned borrowed capacity into a dedicated team of FDEs, deployment strategists and support in about eight months, and how he scopes, staffs and hires for it.
The second issue of the Forward Deployed Engineering newsletter features Justin Sharps, Head of Forward Deployed Engineering at Encord, the data-development platform used for physical and multimodal AI. He describes how Encord moved from borrowing engineers ad hoc for hard customer cases to a named function, organized into pods of engineers, deployment strategists and support, as its customers grew into larger enterprises.
Sharps defines the FDE mandate as engineering for customer outcomes while discounting usual core-product concerns: FDE pods can run their own code paths and own their technical debt, and a workflow that customers actually adopt becomes product validation. Typical engagements include on-premise deployments and custom data integrations, scoped by t-shirt size and outcome rather than hours, with roughly 20% slip treated as goodwill before a formal review.
On hiring, he says many candidates titled FDE turn out to be solutions architects or support, so Encord tests hard system design and data engineering in realistic scenarios. Coding agents have reduced how full-stack a new hire must be on day one, but he still expects engineers to understand fundamentals well enough to vouch for AI-generated code in a critical customer integration. Deployment strategists, by contrast, need customer-facing experience and high agency rather than coding ability.
For FDEs and teams standing up the role, it is a concrete, operator-level account of how a startup defines, scopes and hires for forward deployed work, and a useful reference for interview prep.
Source: www.forwarddeployed.engineering