Forward Deployed Engineering vs. Domain Deployed Engineering™

Forward deployed engineering and Domain Deployed Engineering both place engineers inside a customer’s organization to solve real problems. What separates them is the focus. FDE deploys a technology, while DDE transforms a business process with AI, using a lean squad backed by DataArt’s experience since 1997 and Artisyn, our delivery platform. The model is catching on: by April 2026, FDE job postings on Indeed were up roughly 729% from a year earlier.

Over the past few years, I’ve seen the same pattern with our clients: the pilot works and the demo impresses, but very little changes in how the business operates. The engineering was rarely the problem. What was missing was a team that knew the business well enough to redesign the process around the AI, and stayed until people used it. We built DDE to close that gap.

What Is a Forward Deployed Engineer?

A forward deployed engineer (FDE) is a software engineer who embeds inside a client’s organization to write and ship production code against that client’s real systems and data, rather than handing off a design document and moving on. Palantir Technologies created the role in the early 2010s, calling its FDEs “Deltas”. Its government and enterprise clients held sensitive data in complex, often traditional environments where conventional integration struggled, so Palantir embedded engineers directly with the client and had them build the answer on-site. The Pragmatic Engineer newsletter documents this in detail.

In practice, an FDE splits their time between the client’s site and their employer’s product team. They integrate data, configure and extend the platform, build evaluations, and feed what they learn in the field back into the product. An FDE measures success by the client’s goals but deploys their own company’s platform to get there. That pairing defines the role.

The model spread well beyond Palantir through 2025 and into 2026. Job postings for forward deployed engineers rose roughly 729% year over year by April 2026, according to Indeed data reported by Business Insider. AI companies including Google, Anthropic, and OpenAI got there first, building FDE practices to close the gap between a working AI pilot and a system a customer actually runs in production. Consulting firms followed, and now hire engineers into forward-deployed roles of their own.

What Is Domain Deployed Engineering?

Domain Deployed Engineering (DDE) is DataArt’s AI transformation offering and our evolution of the forward deployed model. Like FDE, it puts engineers inside the client’s organization, building against real systems and data. Its job is to carry a promising pilot to the point where the business actually runs it, which is where most AI initiatives stall. DDE takes specific organizational processes through the full cycle, from ideation through production to adoption and outcome measurement.

DDE combines four elements:

Industry Knowledge × AI Transformation × Engineering Capability × DataArt’s Artisyn

• Industry Knowledge: working fluency in the client’s sector (its workflows and data, and the regulation it operates under), drawn from DataArt’s industry practices in financial services, healthcare and life sciences, travel, media and entertainment, and retail.
• AI Transformation: redesigning the operating model around AI, covering roles, hand-offs, human checkpoints, and adoption, so a model isn’t just bolted onto a process that stays the same.
• Engineering Capability: production-grade agentic engineering, evaluation, and enterprise integration.
• Artisyn: DataArt’s agentic delivery platform. It gives every mission reusable foundations and AI agents, plus governance, in place of a toolchain assembled from scratch. In client engagements, Artisyn has delivered up to 70% faster prototyping and a 30% improvement in development efficiency.

DDE is model-agnostic, even though DataArt partners with Anthropic, OpenAI, and Google Cloud. Those partnerships give DDE teams hands-on experience with each lab’s models, but none of them decides the answer. An FDE employed by a model provider deploys that provider’s models. A DDE mission can recommend whichever model, platform, or combination fits the client’s use case, cost, and risk profile. It tests the options on the client’s own data and changes course as the market moves.

Gartner research finds that “only a little over 40% of selected use cases make it successfully into production” once launched, and separately states that “by 2028, 70% of AI initiatives will underdeliver — not because of model limitations, but because business units lack the capacity to operationalize them.”;That is a problem of how the business runs, so transformation sits at the center of DDE.

How Domain Deployed Engineering Builds on Forward Deployed Engineering

Both models put an engineer close to the problem, writing real code against real systems. They differ in the focus of the work and what the team brings to it.

That doesn’t mean forward deployed engineers lack business fluency. Many build deep expertise in a client’s world over their time on the ground. With DDE, that depth comes from DataArt’s industry practices.

DDE is not our reaction to a trend — it is how we have worked for years, finally given a name and a structure. And we measure it the way clients do: by whether the mission succeeded, not by hours billed.

Scott Rayburn

Chief Marketing Officer, DataArt

Is DDE the Same as Staff Augmentation or Managed Services?

No. Staff augmentation supplies capacity and is measured by hours. Managed services take on ongoing operational ownership. A DDE squad works to a defined mission and answers for the business result, then hands the system back to the client’s own team once the mission succeeds.

The Six Competencies Behind a DDE Mission

To turn the formula into delivery, we built a DDE competency model. Every mission must cover all six, though one person often holds several, which keeps the squad lean.

• Domain and vertical fluency. Working knowledge of the client’s industry, day-to-day workflows, and regulatory context, so each mission targets processes worth transforming.
• Mission craft. Turning an ambiguous situation into a scoped mission: problem discovery, success criteria, sponsor alignment, and a commercial construct that fits.
• Operating-model design. Who does what once an agent handles a step: roles, hand-offs, human-in-the-loop checkpoints, and escalation paths.
• Agentic engineering. Orchestration, tool-calling, retrieval, guardrails, APIs, and authentication wired into the client’s enterprise systems.
• Change and adoption. Getting the client’s people to trust and use the new workflow, and measuring whether they do.
• Evaluation and assurance. Evaluation design, outcomes measured against a pre-mission baseline, and compliance with the safety and regulatory standards of the client’s sector.

How a DDE Mission Runs

DDE work is organized as missions, not open-ended projects. Each one has a fixed end date and follows the same path:

• Charter. Sponsor, success criteria, and commercial construct agreed.
• Discover. Map the real workflow with the business.
• Data readiness. Sources, access, and connectivity confirmed.
• Design. Agents, human checkpoints, and guardrails.
• Prove. Validate on real cases from the client’s business.
• Production. Standards, CI/CD, and sign-off.
• Hand over. Adoption and outcomes measured against the charter’s success criteria; the client’s team runs the system.

Missions run on Artisyn, not on a stack of generic tools assembled for each engagement.

Gartner, Why AI Projects Stall and How to Align Them for Success, Brian Foster and Antonia Roesler, August 17, 2026; and Overcome the AI Value Plateau by Improving Business Readiness for Change, Marie Sienkowski, Kristin Sherwood, and Sneha Ayyar, August 3, 2026. GARTNER is a registered trademark of Gartner, Inc. and/or its affiliates. Gartner does not endorse any vendor, product, or service depicted in its research and does not advise technology users to select only vendors with the highest ratings. This article reflects DataArt’s interpretation of that research; Gartner’s publications consist of opinion and should not be construed as statements of fact, and Gartner disclaims all warranties with respect to them.