FDE Hub
Career infrastructure for Forward Deployed Engineers — the people who take AI from demo to production inside real customer environments.
Where should you start?
Three paths depending on whether you need clarity, interview prep, or open roles.
What is a Forward Deployed Engineer?
FDE meaning, skills, and how the role differs from SWE or Solutions Engineer.
Voices from X
Curated practitioner posts: doctrine, product-loop tests, and title-inflation warnings.
FDE skills roadmap
Interactive curriculum map — adapted from Awesome FDE Roadmap with a 30/90-day plan.
Where are FDE roles hiring?
Curated apply links for labs, hyperscalers, platform, and defense seats.
Why are Forward Deployed Engineers in demand?
Enterprises can buy models; they still struggle to ship them into production under security, data, and workflow constraints. That last-mile gap is the FDE mandate — a pattern Palantir documented years before the agentic-AI hiring wave.
“You can think of a Dev’s focus as “one capability, many customers,” while a Delta’s focus is “one customer, many capabilities.””
Also worth reading CIO on forward-deployed engineering in the agentic AI era, Paraform’s FDSE role guide, and Anthropic on evaluating AI agents.
Start with our plain-language FDE explainer.
What are this week's field signals?
All signals2026-09-18 · FDE Hub
New: Voices from X — curated FDE doctrine, product-loop tests, and title-inflation warningsLenny/Nabeel, Mati’s product loop, the 100-hour test, Gergely’s platform write-back filter, and First Round’s when-not-to hire — packaged as a reading list plus two synthesis guides.
2026-09-14 · OpenAI
OpenAI is posting Forward Deployed Engineer (Legal) seats in SF and NYCMid-September Legal FDE reqs still emphasize embedding with practitioners, shipping a measurable hero workflow, and ~50% travel — domain-shaped FDE tracks are expanding beyond general enterprise.
2026-09-14 · Anthropic
Anthropic Managed Agents add server-side `auto` evaluation for MCP tool callsAs of Sep 10, Managed Agents can run, deny, or pause each agent/MCP call with an evaluation trail — enterprise FDE work is shifting from prompt trust to policy-backed tool authorization.