Field Deployment

Last reviewed: August 2026

A Forward Deployed Engineer (FDE) is embedded directly in a customer’s environment to land the product in production. Rather than design or advisory work, they write and own real production code.

This document summarizes the core knowledge an FDE needs in multicloud environments. It curates CloudPick’s existing documentation from an FDE perspective and adds FDE-specific context.

Origin and Spread of the Role

Palantir created this role in the mid-2000s while deploying Gotham to U.S. intelligence agencies. Classified data, undocumented schemas, and failures of conventional deployment approaches led to a model in which engineers are dispatched directly to the customer site to solve problems.

Since 2024, AI products have faced the same problem, causing the role to spread explosively. For a powerful but general-purpose product (LLMs, agents) to create real value in a customer’s complex legacy environment, someone needs to write code on-site.

Major Events, 2025–2026

Time	Event	Scale
May 2026	OpenAI Deployment Company founded (TPG-led investment)	~$4B
May 2026	Anthropic enterprise services JV “Ode” (Blackstone, Goldman Sachs)	~$1.5B
June 2026	AWS Forward Deployed Engineering organization established	~$1B
2025–2026	FDE job postings up 800%–1,000%+ year over year	39+ companies, 220+ open positions

FDE vs. Solutions Architect

FDEs and SAs (Solutions Architects) both connect customers with technology, but the key difference lies in code ownership and depth of embedding.

Criterion	Solutions Architect (SA)	Forward Deployed Engineer (FDE)
Primary activity timing	Pre-sale – initial implementation	Post-sale – production operations
Core work	Architecture design, technical validation, PoC	Writing production code, integration, live operations
Code ownership	Low (PoC/demo level)	High (directly owns production code)
Customer relationship	Advisory	Embedded
Deliverables	Architecture diagrams, proposals, integration guides	Production code, custom integrations, deployment scripts
Success metrics	Sales conversion rate, platform adoption rate	Deployment success rate, system stability, TTV (Time-to-Value)
Compensation (2026)	$160K–$270K+ TC	$180K–$550K+ TC (frontier AI labs: $1M+)

Who Gets Hired, and When

SA: When product-market fit is clear and technical objections are blocking deal closure
FDE: When the product is powerful but making it work in the customer’s environment requires custom code
SI: When a large-scale system build must be completed on schedule with headcount
Both: For large enterprise deals — the SA designs the architecture, the FDE executes it

FDE vs. SI (System Integrator)

FDEs and SIs (System Integrators) both write code on-site at the customer, but they differ in purpose and ownership structure.

Criterion	FDE	SI (System Integrator)
Affiliation	Product company (model company, SaaS vendor)	Separate consulting/SI firm
Purpose	Land the company’s own product in the customer’s environment and feed product feedback back to headquarters	Build and deliver a system within the SOW (statement of work) scope on schedule
Code ownership	Can contribute to the product core; reusable patterns feed back into the product	Owned by the customer; transitions to a maintenance contract after project completion
Duration	Until the product lands (weeks to months, recurring)	SOW duration (months to years, clearly ending)
Success metrics	Product adoption rate, TTV, contribution to product improvement	Meeting deadlines, passing acceptance tests, sign-off
Staffing model	Small elite team (1–3 people)	Large-scale staffing (dozens to hundreds)
Playbook	None (works where a playbook doesn’t yet exist)	Exists (methodology, deliverable templates)
Product feedback	Core role — reflects field problems into the product roadmap	Limited — separate organization from the vendor

Core Knowledge FDEs Need

Week 1: Entering the Customer Environment

The realities you encounter when first entering a customer environment. This is an FDE-specific challenge that SAs or typical SWEs don’t experience.

Challenge	Description	Related documentation
Working in someone else’s account	Operating with least privilege in the customer’s AWS/Azure/GCP account	Account and Organization Structure
Negotiated IAM	Restricted access permissions coordinated with the customer’s security team	IAM Overview
Air-gapped environments	Air gaps, proxy-only access, restricted internet access	Network Isolation
Remote access constraints	Customer-specific VPNs, bastion hosts, zero-trust access	Remote Access Management
Compliance	Industry-specific regulations of the customer (finance: PCI-DSS, healthcare: HIPAA, public sector: FedRAMP)	Compliance

Infrastructure and Cloud

An FDE must be able to work with whichever cloud the customer environment uses.

Getting Started with Cloud — basic concepts
Comparing Vendors — vendor differences for equivalent functionality
Regions and Availability Zones — data sovereignty, latency
Shared Responsibility Model — the boundary of responsibility in a customer environment

Security and Governance

The customer’s security team is often an FDE’s toughest stakeholder.

IAM in Depth — cross-account access, temporary credentials
Secrets Management — accessing secrets in a customer environment
Data Protection — handling customer data
Zero Trust — modern enterprise access model
Landing Zone — understanding the customer’s cloud foundation

Building and Operating

If an FDE owns production code, they also need to understand operations.

Getting Started with DevOps — CI/CD, automation
Monitoring → Observability — debugging in the customer environment
SLI/SLO — defining success metrics
Disaster Recovery — responding to failures in the customer environment
Security Incident Response — collaborating with the customer during an incident

Data and Integration

Understanding and connecting the customer’s data is at the core of FDE work.

Database Operations — accessing and integrating with the customer’s DB
Database Migration — migrating data

AI/Agent Deployment

The direct driver behind the surge in FDE demand in 2025–2026.

Getting Started with AI — understanding AI products
Advanced RAG Patterns — building RAG on customer data
AI Agents — deploying agentic workflows
LLMOps — evaluation, cost, operations, agent observability
AI Security — guardrails, prompt injection defense

The FDE Tech Stack

Every customer environment differs, but these are the tools FDEs commonly work with.

Category	Tools
Languages	Python, TypeScript, SQL, Go/Java, Bash
Backend	FastAPI, NestJS, Spring Boot
Frontend	React, Next.js (ops dashboards, customer portals)
Data	PostgreSQL, Spark, dbt, Airflow, Kafka
Cloud	AWS, Azure, GCP (EC2, K8s, Lambda, S3, etc.)
Containers	Docker, Kubernetes, Helm
IaC	Terraform, Pulumi, CloudFormation
CI/CD	GitHub Actions, GitLab CI, Argo CD
Observability	Datadog, Grafana, Prometheus, Sentry
AI/agents	LangGraph, CrewAI, Semantic Kernel, vector DBs

The Shift to Agentic Delivery

As of 2026, how FDEs work is changing.

Before: Writing glue code by hand on-site, manual integration, taking weeks to months

Now: A small human team combined with AI agents

Agents handle scaffolding, evaluation, and long-running workflows
Humans (FDEs) focus on discovery, governance, and go-live decisions
AWS’s AI-Driven Development Lifecycle formalizes this pattern

What doesn’t change: Resolving ambiguity in the customer environment, persuading stakeholders, and production ownership — these are not automated.

Career Perspective

Path	Description
IC track	FDE → Senior → Staff/Principal FDE
Product transition	Field feedback → PM/TPM (the person who best understands product gaps)
GTM leadership	Leading a technical sales/deployment organization
Founding a company	Many founders are former Palantir FDEs — the position that understands customer problems most deeply