Forward Deployed Engineering

I spent 7 years at Palantir as a Forward Deployed Engineer and hiring manager. The role is having a moment — Anthropic and OpenAI are pouring billions into it, founders are spinning up "FDE" titles before they've worked out what the role does — and most of what gets written about it is muddled.

This page is the front door to everything I have on the subject: three essays that make the argument, two on training FDEs, the framework that operationalises it, the open materials, case notes and how to work with me.

The thesis

If LEGO Had Forward Deployed Engineers What actually separates an FDE from a really good consultant. The difference is the loop: FDEs build the tools that bootstrap themselves, and the best of those tools walk back to product. FDE work is product R&D dressed as delivery.

The FDE Fork: Platform or Outcomes AI made a second business model viable. You can productise the weird bricks and sell a platform (Path A, the Palantir-to-Foundry path), or keep the tooling internal and sell outcomes (Path B). Both are coherent. Founders who don't consciously pick one build a confused org with a confused FDE role.

Is FDE a Bubble? Palantir never chose FDE; the role was residue left by a decade of solving problems from first principles. What is spreading now is the title, not the reasoning. The question is not whether FDE is a bubble but whether you derived your structure from your deployment problem.

Also: What Claude Code's Creator Validated About FDE (outside confirmation of the thesis) and Forward Deployed Engineers and how to hire them (the original 2024 piece on the role).

Training FDEs

The labs teach the tools. The hard part is the craft of making a deployment work: finding the person who makes the call, building around the decision, and getting the work used. That grows through feedback on real work from people who have done it.

How Do You Grow FDEs? Most FDEs have to be grown. Palantir threw people in at the deep end, but that worked because of the mentors, strategists and product team around them. Most companies don't have that, so their training has to fit their own version of the role.

The Fast Layer and the Slow Layer An FDE's skills come in two layers. The tools change every few months. The craft of making a deployment work barely changes. Select for the traits, train the stack, and teach the craft through real work and fast feedback.

The framework

The Platypus FDE Blueprint — the essays turned into a ruler, published free under CC BY. Nine parts: a reasoning gate for whether you need FDEs at all; the fork; five maturity stages that set the field's autonomy; the two-founder unit; the three interfaces with product, sales and steady state; two KPIs and one honest proxy; the hiring system; the eight ways it fails; and how it is applied to a real company.

Case notes

The Platypus case notes cover four different shapes: building an FDE unit out of a solutions team, standing up the motion inside a large services firm, increasing impact per FDE in a company that already has the model, and diagnosing a field team whose problems were elsewhere in the organisation.

Open materials

FDE Advisory Materials — templates and open materials: hiring loop and scorecards, the product–FDE interface agreement, technical enablement standards, operating models, job descriptions, and an onboarding checklist. Free to use.

Work with me

I run Platypus Technologies. The FDE Programme gets new FDEs ready to lead deployments. It is not a course. It runs for eight weeks, remotely. Each FDE starts with a two-day exercise on a realistic client case. Then a coach reviews their own live deployment every week for six weeks. At the end, each FDE defends a fresh case to a panel and gets a written verdict on whether they are ready. There is a team programme for startups with a few new FDEs, and a cohort programme for services firms building a practice.

Organisational work starts with the FDE Readiness Assessment: a two-week review and a written readout of where you sit against the Blueprint and what to change.