What Is Forward Deployed Engineering? The Role Behind the 729% AI Hiring Surge

Forward Deployed Engineers (FDEs) are suddenly everywhere. OpenAI is building a company around them. Anthropic is investing $100 million in training them. And according to Indeed data reported by Business Insider, job postings for FDEs grew by 729% in just one year, from 643 in April 2025 to 5,330 in April 2026.

There’s a practical reason for the sudden interest. Companies have access to powerful AI models, cloud platforms, and development tools. Yet getting software to work with their existing systems, data, and daily operations remains difficult. A successful demo can be built in days. Making it reliable enough for everyday use can take considerably longer.

Forward Deployed Engineering brings engineers directly into that work. They understand the client’s business, make technical decisions, write code, and stay involved as the solution reaches production.

Let’s look at what the role involves, why it’s gaining traction, and what companies can expect from this approach.

What Is a Forward Deployed Engineer?

A Forward Deployed Engineer (FDE) is a software engineer who works directly with customers to understand their business problems, build solutions, and get them running in production.

Unlike engineers who primarily work on a company’s core product, FDEs spend much of their time working with individual customers. They learn how the business operates, examine existing systems, make technical decisions, write code, build integrations, and solve problems that arise during deployment.

The role became widely known through Palantir, which has long employed Forward Deployed Software Engineers (FDSEs), internally called Deltas. Palantir describes the difference between its product engineers and FDEs in two short phrases:

• Product engineers: “One capability, many customers.”
• Forward deployed engineers: “One customer, many capabilities.”

A product engineer might spend months improving a feature used by thousands of customers. An FDE works across different parts of a system to solve problems for a particular customer. That could mean connecting legacy applications, changing data flows, building custom features, or adapting existing software to fit the customer’s operations.

The exact responsibilities depend on the company. At OpenAI and Anthropic, FDEs help customers integrate AI into their systems and workflows. At Palantir, they work with customers to build and deploy applications around complex data and operational needs.

What these roles share is a combination of strong engineering skills, direct customer involvement, and ownership of technical delivery.

What Does a Forward Deployed Engineer Actually Do?

Consider a company that wants an AI assistant to help its support team resolve customer requests. Sounds relatively simple. Connect an AI model to company documents, let employees ask questions, and produce useful answers.

Once engineering begins, the team discovers several complications. Customer records sit across different systems. Some documents are outdated. Employees have different access rights. Certain actions require a manager’s approval. And the assistant must know when to hand a request to a human.

An FDE works through these problems with the people who understand them.

A typical engagement involves:

• Learning how the business works. The engineer talks with users and stakeholders, examines existing workflows, and identifies the systems and data involved.
• Defining the technical approach. They decide what can be reused, what needs to be built, how systems should connect, and which risks must be addressed.
• Building the solution. They write production code, develop integrations, set up data flows, and work with the client’s technical team.
• Testing in the real environment. They check whether the software handles actual business data, user permissions, exceptions, and performance requirements.
• Improving what reaches production. They resolve issues, respond to user feedback, document decisions, and help the client’s team maintain the solution.

These activities can overlap. An engineer may discover a problem with the original design while building an integration, discuss it with the customer that afternoon, and change the implementation before the next release. FDEs also need the judgment to challenge a proposed solution. If an expensive AI workflow could be replaced with a simple rules-based system, they should be able to explain why that approach makes more sense.

How Is an FDE Different From a Software Engineer or Technical Consultant?

The roles share many skills. A traditional software engineer may work directly with customers, while an experienced technical consultant may write production code. The distinction usually comes down to the scope of responsibility and how closely the engineer works with the customer.

Role	Typical focus
Software Engineer	Develops, tests, and maintains software within an established product or project team
Forward Deployed Engineer	Works directly with the customer, helps define the technical approach, builds the solution, and supports its deployment
Solutions Engineer	Helps customers understand and adopt a technical product, often during evaluation, sales, or initial implementation
Technical Consultant	Assesses technical challenges, recommends approaches, and may help implement them, depending on the engagement
Staff Augmentation Engineer	Joins an existing team to provide engineering capacity and carries out responsibilities agreed upon with the client

An engineer working through staff augmentation could take on forward-deployed responsibilities if the project requires direct work with stakeholders, broad technical ownership, and involvement from planning through deployment.

Why Is Forward Deployed Engineering Growing So Fast?

Forward-deployed work existed long before the current AI boom. Complex enterprise systems have always required engineers who understand customer environments. AI has made that need much more visible.

1. Companies Are Running Into the Limits of AI Prototypes

Building a promising AI prototype has become relatively easy. Deploying one inside a large organization is another matter. We’ll drive our example from a systematic review of generative AI in manufacturing and industrial contexts published in MDPI’s Electronics.

They analyzed 252 peer-reviewed studies published between 2022 and June 2026 and found that:

• 84.5% of the studies were classified as research-stage.
• 14.7% reported pilot or case-level evidence.
• Only 0.8% indicated operational industrial evidence in their abstracts.

These figures describe the maturity of published research, rather than the success rate of commercial AI projects. Still, they show how much work remains between demonstrating a technology and proving its value in industrial operations.

The researchers identified several needs for real-world adoption, including domain-specific data, verification, cybersecurity, system integration, and human accountability. These are exactly the types of problems a forward-deployed engineer may encounter.

2. AI Is Moving Into Everyday Business Workflows

In its 2025 State of Enterprise AI report, OpenAI reported a 19-fold increase in weekly users of Custom GPTs and Projects during 2025. Its September 2025 Economic Index found that 77% of business API use followed automation patterns, compared with approximately 50% of Claude.ai use.

So, the the way companies use AI is changing. Companies are going beyond occasional AI-assisted tasks and building repeatable workflows around the technology.

Using AI through a chat interface places much of the work in the user’s hands. Building an automated workflow requires software that can retrieve data, follow business rules, interact with other applications, handle failures, and sometimes act without a person approving every step. And guess who has to design and build these workflows? A forward deployed engineer.

3. Software Teams Still Struggle to Understand Customer Requirements

The case for FDE extends well beyond AI. A 2023 MDPI study on requirements engineering surveyed 84 practitioners about how they gather, analyze, and manage software requirements.

The researchers found a few recurring problems: misunderstood requirements, errors and omissions, the effort required to document requirements, and gaps in customers’ technical knowledge.

These problems are common in software development:

• Business stakeholders know what they need but may overlook technical constraints.
• Engineers understand the requirements but may miss the business context behind them.
• Written specifications can’t capture every exception, workflow detail, or change in business needs.

Forward-deployed engineers help close these gaps by working directly with the people who use the software. They can examine workflows, clarify requirements, test ideas, and adjust the solution as new needs emerge.

What Value Does a Forward Deployed Engineer Bring to a Project?

Closer collaboration sounds useful, but companies need a stronger reason to invest in it. The real value of FDE depends on what changes during delivery and what the customer gains from those changes.

Fewer Costly Assumptions

Consider an engineer asked to develop a reporting dashboard. The specification lists several metrics, filters, and charts. During discussions with the operations team, however, the engineer discovers that employees spend hours correcting the underlying data before they can produce their reports.

Instead of building another dashboard, the engineer can address the actual problem:

• Fix data quality issues at the source
• Improve data flows between systems
• Add validation rules to reduce manual corrections

This kind of early technical involvement can help reduce rework, particularly when the business problem is still being defined.

Faster Technical Decisions

In conventional delivery, a technical question can pass through several people before reaching someone who can answer it. A developer raises an issue. A project manager forwards it. A business stakeholder reviews the request. A decision returns to the engineering team days later.

An FDE works directly with stakeholders and can often resolve issues much sooner. This means:

• Fewer rounds of clarification between business and engineering teams
• Quicker decisions when requirements or priorities change
• Less time spent waiting for answers during development

Software That Fits Existing Operations

Enterprise software isn’t operating in vacuum. A new product may depend on legacy databases, internal APIs, third-party tools, user permissions, and workflows that have evolved over many years. Forward-deployed engineers study these dependencies while designing the solution. or example, when building an AI agent that uses CRM records, ERP data, and internal documents, the engineer needs to:

• Connect the relevant systems and manage data access
• Define what the agent can do and which actions need human approval
• Handle errors, missing data, and unusual cases
• Test the agent against real business workflows

Clearer Ownership of Technical Problems

Complex projects often produce issues that fall between teams. An integration fails even though both connected systems work. A release causes unexpected performance problems. A feature works for one department but disrupts another.

An FDE can step in to:

• Investigate problems across systems and teams
• Identify the root cause and work out a fix
• Coordinate changes with the people responsible for each system
• Follow the issue through testing and deployment

The client has an engineer who understands the broader technical picture and can take responsibility for resolving problems.

Earlier Feedback and Better Long-Term Handover

A customer may only discover that a workflow is inconvenient after employees start using the software. Forward-deployed engineers can work with those users during rollout to:

• Gather feedback directly from users
• Fix issues and improve workflows based on real usage
• Document technical decisions and system behavior
• Help internal engineers understand and maintain the solution

This helps the customer’s team take over with a clearer understanding of how the software works and what it needs to keep running.

What Does Forward Deployed Engineering Look Like in Practice?

The value becomes easier to understand through real engineering work.

Palantir: Reducing AI Costs at Tampa General Hospital

At Tampa General Hospital, a Palantir Forward Deployed Engineer worked with healthcare experts to improve an AI-driven utilization review workflow. The team reviewed model options, refined prompts, and replaced AI calls with conventional code where it made sense.

According to Palantir’s published DevCon case, the work produced:

• Nearly 70% lower costs for AI-generated utilization reviews.
• 84% fewer GPT calls.
• A preference for the optimized output among the hospital’s experts in 90% of comparisons.

Exoft: Turning Sales Expertise Into an AI Product

Sometimes a client knows the business problem well but needs a technical partner to turn that knowledge into software. That was the starting point for Exoft’s work with a US sales technology startup.

The founders had extensive experience in value-based selling and wanted to create an AI-driven sales platform that would help teams improve discovery, qualification, and value propositions. Exoft became the client’s exclusive technology partner.

The team developed the product around real sales workflows, including AI-assisted analysis of discovery conversations, document creation, deal scoring, and CRM-related workflows. This project illustrates another use of forward-deployed practices: working with people who hold deep business knowledge and helping turn that expertise into a product customers can use.

When Does Your Company Need a Forward Deployed Engineer?

Some projects have stable requirements, an established architecture, and a clear delivery plan. An experienced software development team may be all they need.

Forward-deployed engineering becomes particularly useful when technical work depends on frequent decisions with the customer’s team.

Consider the following situations:

When You Need an FDE	How They Can Help
You have an AI use case but don’t know how to deploy it.	Connect business data, manage permissions, test AI outputs, and turn a working prototype into a reliable production system.
Your project depends on several existing systems.	Integrate ERP, CRM, legacy applications, and other tools while working directly with the people who understand them.
Your requirements are still taking shape.	Help define the scope, test ideas, make technical decisions, and adjust the solution as business needs become clearer.
Your internal team needs senior technical ownership.	Take responsibility for architecture, integrations, and complex technical decisions while working alongside your engineers.
You need to improve software that’s already running.	Investigate performance issues, fix unreliable integrations, and improve workflows based on feedback from real users.

In each case, the right setup depends on the complexity of the work and the skills already available within the company. A dedicated FDE may be appropriate for one critical workstream. A larger project may need several engineers supported by architects, QA specialists, DevOps engineers, and product experts.

Conclusion: Forward Deployed Engineering at Exoft

At Exoft, we work with clients whose software needs rarely fit into a perfectly defined task list. Some have a product idea and need help deciding what to build. Others have existing platforms that must be modernized without disrupting daily operations. Many need to connect new technology with systems their teams have used for years.

Our Forward Deployed Engineering approach brings technical expertise into the customer’s work early and keeps it involved as the solution develops. We work alongside your team to understand the problem, examine the systems behind it, choose a practical technical approach, and build software that can run in your environment.

If your team is working through a complex build, integration, or AI deployment, talk to Exoft. We can help you understand where forward-deployed engineers would fit and what they could take ownership of.