Forward deployed engineering began with on-site deployment and platform stability, then expanded into data integration, ontology work, custom solutions, and customer enablement.
2
The role has no stable definition because each company has inherited a different version of it, but customer accountability connects the different jobs.
3
As coding agents make software cheaper to produce, engineering work moves closer to customers and outcomes, making agent engineering a newer form of forward deployed engineering.
Summary
Natalie Meurer traces forward deployed engineering from its origins at Palantir. In the early years, the job meant sitting with customers, often at on-premise deployments, and keeping the platform running. It later included data integration, modeling customer data through an ontology, building solutions in Slate and Foundry, and teaching customers to do more of this work themselves. The title expanded as the work expanded, so companies now use it for different roles and hiring expectations. Meurer argues that customer accountability is the common thread. Coding agents make it faster to build software, which gives engineers more room to work directly on customer problems and product changes. She connects this shift to outcome-based pricing. When a company is paid for usage or a result, someone must be accountable for delivering that result. In her view, agent engineering is one form of forward deployed engineering, and more engineering roles are moving in that direction.
Forward deployed engineering began with physical deployment at the customer
Palantir job postings in 2016 emphasized the word "forward" because engineers were expected to sit with customers and work on the ground. This was partly driven by on-premise deployments. Meurer describes the early role as something close to DevOps rather than the broad customer-facing engineering job associated with the title today. Her own onboarding project was to deploy the software on an EC2 instance. The work could include responding to operational problems at night, such as a customer accidentally unplugging the instance. Platform stability made up much of the early job.
The role expanded from keeping systems alive to integrating customer data
Once the platform became more stable around 2012, the harder problem was getting useful data into it. Meurer compares data integration software without integrated data to a movie theater with no movies playing. Customer data lived in many places, so forward deployed engineers had to connect those sources and understand the customer's environment. They also modeled the data through what Palantir called an ontology, which Meurer describes as a taxonomy for the data. The job therefore combined DevOps with data integration and required engineers to understand how the customer's information should be organized.
Custom solutions and enablement added more work to the role
By 2016, forward deployed engineers were building custom solutions to known problems. In Palantir's Slate, a drag-and-drop builder, they mapped interface components to data sources so customers could use the integrated data. This work later gave way to Foundry, whose focus was turning data into decision-making. Meurer says dashboards became less useful when they did not write back to the data source, because they decayed over time. Around 2020, Palantir also wanted customers to do more of this work themselves. Forward deployed engineers therefore added customer enablement to DevOps, data integration, and solution building.
The title covers different jobs because each company has a different FDE vintage
Meurer says forward deployed engineering is not one thing. The tasks accumulated instead of replacing one another, so the total set of work grew over time. She describes the role as a strong training ground for generalists because it exposes people to platform stability, data work, custom solutions, and customer enablement. She suggests asking candidates which "vintage" of FDE they are. The 2008 version focused on platform stability and panic. Later versions added data integration, custom solutions, and enablement. Current job descriptions can demand staff-level engineering experience alongside direct sales, solution architecture, and teaching experience.
Customer accountability is the thread connecting the different versions
Even though the tasks differ, Meurer sees one continuity point: forward deployed engineers are accountable to the customer. That accountability can apply to DevOps, enablement, custom solutioning, or data integration. She uses this idea to explain why the role still matters despite its unstable definition. The common element is responsibility for solving a customer's problem rather than ownership of one narrow technical activity. This also explains why the role has spread across companies under names such as forward deployed, customer, or deployed engineering.
Cheap code makes customer and product work converge
Meurer asks what happens to engineering when a prompt to a coding agent can quickly produce something useful. Forward deployed engineers can now talk to customers, prototype, and build end-to-end solutions with coding agents. At Sierra, she says the boundaries are also blurring in the other direction. Product engineering is becoming more client-facing, and product engineers should think about the customer and the product together. Infrastructure engineers also need to consider how software is deployed. As code becomes cheaper to produce, understanding the customer and carrying responsibility for the result become more important parts of engineering work.
Outcome-based pricing creates a need for engineers who own the result
Meurer connects the evolution of engineering roles to software pricing. Seat-based pricing fits cases where it is difficult to attribute an outcome to the product. Usage-based pricing, such as paying foundation model providers, charges for consumption when the outcome is harder to attribute. Customer experience systems can be priced by outcome, such as making a sale or solving a customer inquiry. Meurer expects more pricing to move toward outcomes. When a company charges for a result, it needs a way to deliver and guarantee that result. She identifies forward deployed engineering with that responsibility.
Agent engineering is a newer form of forward deployed engineering
Meurer originally described agent engineering as a subdiscipline of AI engineering with the same customer accountability as forward deployed engineering. She now thinks that framing was incomplete. Agent engineering is a flavor of forward deployed engineering, alongside product, AI, solutions, and customer engineering. The shared work includes enablement, DevOps, solution building, data integration, creating an agent, and deploying it into production for a customer. Her closing claim is that the term forward deployed engineering may lack a coherent definition, while the work it points to is spreading across engineering.
"The one continuity point is that you actually have every single forward deployed engineer accountable to the customer."11:57
Who should watch
You are hiring forward deployed, customer, or deployed engineers and need to decide what work the title should cover.
Your engineering team is using coding agents and needs to keep customer understanding and delivery responsibility in the role.
You are pricing an AI product by usage or outcomes and need engineers who can connect the product to a customer result.
