Palantir FDE Model: Deep Dive

The definitive analysis of Palantir's Forward Deployed Engineering model and its application to DaisyAI.

What is an FDE?

A Forward Deployed Engineer is a software engineer who alternates between being embedded with customer teams and core product engineering teams. They're on-site with customers (sometimes for months) using the company's tech stack to build production-ready workflows that solve specific client problems.

"FDE responsibilities look similar to those of a startup CTO: you'll work in small teams and own end-to-end execution of high-stakes projects." — Palantir

Key distinction: FDEs aren't just consultants or sales engineers. They:

Write production code directly on customer infrastructure
Use live production tooling (not anonymized data)
Maintain long-term customer relationships
Contribute insights back to core product development

Origin Story: The French Restaurant Analogy

Ted Mabrey (Palantir's head of commercial) explains the FDE role was inspired by how excellent French restaurants operate: the waitstaff is an intrinsic part of the kitchen, empowered to tell customers "no" if they're ordering the wrong wine with fish.

Palantir recognized that enterprise software needed engineers who would take ownership of customer outcomes, even when customers didn't know how to ask for what they needed. This meant breaking down traditional barriers between sales, implementation, and engineering.

Palantir's Organizational Model

Historical Scale

Until 2016, Palantir had more FDEs (called "Deltas") than software engineers
Created the role in early 2010s for government and private clients
Later launched Palantir Foundry, allowing FDEs to transition into core product roles

Structure

FDEs report to Business Development
Work in small teams supporting individual customers
Expect ~25% onsite time at customer locations
Some manage direct reports across remote offices

Platform Strategy

What makes Palantir effective is they approached this as a platform product company:

Each FDE builds prototypes using platform services
Platform product org continuously generalizes learnings
Creates virtuous cycle: customer work improves underlying product

The Economics

Contract Size Requirements

The FDE model only works at scale:

Minimum viable: 7-figure contracts ($1M+)
Ideal range: 8-9 figure contracts
Anti-pattern: Applying FDE model to $20K-$100K contracts is "a deeply misguided exercise in capital allocation"

"The power of the FDE model is intrinsically linked to the complexity and scale of the problem it's solving. You've got to be solving a hard enough problem."

Revenue Model

Services drive product adoption, not standalone revenue
Palantir funds engineering time upfront to land customers
Over time, revenue mix tilts toward software subscription
Success metric: Revenue per FDE should rise as software replaces manual work

Warning Sign

If revenue per FDE flatlines and custom work keeps accumulating, you may be running a services shop, not a product company.

What Makes a Successful FDE

Background

Solid software engineering fundamentals
Real-world project experience (shipped products)
Range: 1+ years (Palantir) to 5+ years (senior roles at Ramp)

Key Traits

Comfort with ambiguity - undefined problem spaces are the norm
Strategic thinking - "startup CTO" mentality
Boundary-setting - knowing when to say no to meetings
Domain expertise - deep understanding of customer's industry
Customer charisma - can build trust and long-term relationships

The Duality Challenge

FDEs must balance two distinct skill sets:

Customer consulting (understanding needs, building trust)
Platform engineering (writing production code, improving product)

This tension requires discipline. FDEs must actively limit non-productive activities.

Companies Using FDEs (2024-2025)

Company	Industry	FDE Team Size	Notes
Palantir	Enterprise Software	Large	Originated the model
OpenAI	AI	10+ across 8 cities	Established early 2025
Ramp	Fintech	~15 in pods	9 months experience
Salesforce	CRM (Agentforce)	Unknown	Enterprise division
Commure	Healthcare AI	Unknown	Direct competitor
Matta	Industrial AI	Unknown
Gecko Robotics	Robotics	Unknown
Lindy	AI Agents	Unknown

FDE vs. Related Roles

Role	Key Difference
Solutions Architect	Advisory only, rarely writes production code, uses anonymized data
Consultant	One-off recommendations, no product contribution
Sales Engineer	Focused on closing deals, not implementation
FDE	Long-term relationships, production code, feeds back to product

Why AI Specifically Needs FDEs

The Implementation Gap

95% of AI projects fail to create business value (MIT 2024)
Only 1% of companies have reached full AI maturity
74% of companies struggle to scale AI value
FDE demand is skyrocketing as implementations become the bottleneck

The Process Reality Problem

"Standard operating procedures are corporate fiction: static, incomplete, and often wildly outdated. AI systems need behavioral fidelity—how the work actually gets done—not the idealized version."

FDEs solve this by:

Discovery - Shadowing users to reconstruct actual processes
Integration - Wiring AI into existing tools (ERPs, CRMs, call centers)
Trust - Building relationships that reveal process nuance

The New Moat

"Enterprise AI isn't a static deployment. It's a continuous loop between humans and systems. The companies winning aren't just training better models—they're building better relationships with their customers."

The moat isn't proprietary models—it's proprietary implementation.

Lessons for DaisyAI

What We Should Adopt

Platform mindset - Every engagement improves the core product
Contract selectivity - Target 6-7 figure opportunities minimum
Hiring bar - Top-tier engineers with customer charisma
Long-term relationships - Not one-off consulting gigs
Domain focus - Healthcare expertise is our edge

What We Should Adapt

Scale - We're 2 people, not Palantir; start with 1-2 deep engagements
Product foundation - We have working SaaS; use it as proof point
Network leverage - Wharton connections for warm intros
Free pilots - Build case studies before charging premium rates

Warning Signs to Avoid

Revenue per engagement flatlines
Custom work doesn't feed back into product
Spending time on sub-$100K opportunities
Becoming a body shop instead of platform company

Sources

Last updated: 2025-01-07