What a Forward-Deployed Engineer Actually Does at Palantir, Runway, and Greptile (April 2026)
John Kim
•
Co-founder @ Paraform
April 22, 2026
April 22, 2026
Most people think they understand what a forward deployed engineer does, but they're picturing a solutions engineer who codes sometimes. That's not it. FDEs write production-grade software while embedded directly with the customer, owning everything from technical discovery to post-deployment fixes. Palantir invented the role because their software was too complex for customers to just plug in, and now the model has spread to AI startups where one engineer owns an entire customer relationship from first call to production.
A forward deployed engineer sits at the intersection of software engineering and customer delivery. The job? Write production-grade code, but do it while embedded directly with the customer. FDEs don't build features in a vacuum - they ship solutions on-site, translating a product's raw capability into something that actually solves a specific business problem.
Palantir coined the role over a decade ago out of necessity. Their software was powerful but complex, and customers in government and defense couldn't just plug it in. Someone needed to sit in the room, understand the mission, and build integrations that worked under real constraints. That someone became the forward deployed engineer.
The FDE role exists because great software alone isn't enough. Someone has to close the gap between what a product can do and what a customer needs it to do, in their environment, on their timeline.
Since then, the model has spread well beyond Palantir. AI startups like Runway and Greptile now run their own versions of the role, adapting it for faster iteration cycles and smaller teams. The core premise remains the same: put an engineer where the problem lives.
Palantir calls its FDEs "Deltas," and until 2016, the company had more Deltas than traditional software engineers. That ratio tells you everything about how central customer-facing engineering was to Palantir's growth.
The deployment model works like this: a Delta embeds directly with a customer team in government, financial services, or commercial operations. Their job is to take Palantir's core products, Foundry and Gotham, and make them functional inside environments that are messy, siloed, and often classified. That means:
Deltas don't hand off a spec to a product team and wait. They own the technical outcome from start to finish, sitting shoulder to shoulder with the people using the software.
At a company like Greptile, which builds AI-powered code review tools, the FDE role looks different from Palantir's large-scale government deployments. The problems are smaller in scope but move faster. An FDE at an AI startup might spend Monday running technical discovery with a prospective enterprise customer, Tuesday building a proof-of-concept tailored to that customer's codebase, and Wednesday walking their security team through compliance requirements. This role is part of the expanding range of startup engineering positions companies need to fill in 2025.
The day-to-day typically includes:
Where Palantir's FDEs operate inside massive institutional environments with years-long contracts, startup FDEs work in compressed cycles. A proof-of-concept might need to land in a week while the customer compares three competitors simultaneously. Speed and technical depth have to coexist.
FDE job postings soared by more than 800% between January and September of 2025, according to data tracked across major hiring sites. That surge reflects how quickly AI startups have adopted the model Palantir built, reshaping it for smaller teams where one engineer owns an entire customer relationship from first call to production deployment.
The titles sound similar, but the work diverges sharply. A solutions engineer typically supports the sales cycle: running demos, answering technical objections, and designing proposed architectures. Once the deal closes, they move on. A sales engineer operates in a similar orbit, acting as the technical voice during pre-sale conversations. Neither role expects you to ship code that runs in production.
FDEs do. They write production code, deploy it inside the customer's environment, and own the technical outcome well after the contract is signed. There's no handoff to an implementation team. If something breaks at 2 AM in a customer's pipeline, the FDE is the one fixing it.
| FDE | Solutions Engineer | Sales Engineer | |
|---|---|---|---|
| Writes production code | Yes | Rarely | No |
| Owns post-sale delivery | Yes | No | No |
| Embedded with customer | Long-term | During sales cycle | During sales cycle |
| Reports into | Engineering | Sales or pre-sales | Sales |
If you're assessing candidates or considering the role yourself, the question to ask is simple: does this person own the outcome after the deal closes, or before it? That's the dividing line.
The technical foundation hasn't changed much: Python, TypeScript, SQL, and fluency with cloud infrastructure like AWS or GCP. Container orchestration with Docker and Kubernetes remains table stakes for anyone deploying inside customer environments. That's the deep end of the T-shape.
The broad end is where most candidates fall short. FDEs need customer empathy, clear communication under pressure, sharp problem decomposition, and genuine product sense. Similar skills appear in design engineering roles, which command compensation premiums. You're writing code, yes, but you're also deciding what to build, for whom, and why it matters right now.
For AI-focused roles, the bar has shifted. Agentic orchestration, evaluation frameworks for model output, and AI observability tooling are increasingly listed as requirements. If you can't reason about why a pipeline is hallucinating in production, you're not ready for an AI startup's FDE seat.
Compensation reflects the hybrid demand. The median FDE salary sits at $173,816, and 70% of FDE postings mention equity on top of base pay.
At Palantir, average total compensation for an FDE lands around $238,000, with the range stretching from $205,000 to $486,000 depending on level. Staff-level FDEs clear $630,000 or more according to compensation data tracked across the industry.
Why does this role pay a premium? The talent profile is genuinely rare. You need someone who can write production code, handle ambiguous customer requirements, and make high-stakes technical decisions without a product manager in the room. Most engineers can do one of those things well. Finding someone who does all three is a different search entirely.
For companies benchmarking FDE offers, the takeaway is straightforward: if you're posting comp bands that look like a standard software engineering role, you'll lose candidates to organizations that understand what they're actually buying. This is just one aspect of effective engineering hiring strategies that matter in 2024. The client-facing dimension commands its own premium, and the market has priced that in.
FDEs sit in a hiring dead zone. You need someone who can ship production code, run a customer meeting, and make architectural calls on the fly. Hiring software engineers with this unique blend of skills requires a different approach. That combination is rare, and traditional recruiting methods aren't built to screen for it. Job boards surface engineers. Agencies surface salespeople. Neither pipeline reliably produces the hybrid profile an FDE role demands.
That's where Paraform comes in. Our network of specialized recruiters understands roles like this because they've filled them for companies like Palantir, Rippling, and Cognition - Windsurf, for example, made seven FDE hires through Paraform in just a few months.
When a recruiter on our network has spent years placing engineers into client-facing technical roles, they know what to look for beyond the resume. This specialized approach is part of modern tech talent recruiting strategies that actually work. As one analysis of 1,000 FDE job postings found, the requirements vary wildly across companies, which makes specialized recruiting knowledge even more valuable. With over 1,000 customers, $50M+ paid to recruiters, and ~12 days to meet the hire, Paraform has the infrastructure to match hard-to-fill roles with people who actually close them.
The forward deployed engineer Palantir created over a decade ago has become the hiring template for AI companies that need someone to own the entire technical relationship with a customer. You're looking for an engineer who ships production code, runs customer meetings, and makes architectural calls without waiting for a product team to weigh in. That combination is rare enough that traditional recruiting pipelines consistently miss it. If you're trying to fill one of these roles, request a demo to work with recruiters who've placed FDEs at companies like Palantir and Cognition. The companies building serious AI products aren't posting on job boards and hoping - they're working with people who know where this talent actually lives.
FDEs write and own production code deployed in customer environments and stay engaged long after the deal closes, while solutions engineers focus on pre-sale technical support and hand off implementation to other teams. The dividing line is simple: FDEs own the outcome after the contract is signed, not before it.
No. The role demands screening for a rare hybrid profile: production coding ability, customer empathy, and autonomous decision-making under pressure. Most job boards surface engineers, and most agencies surface salespeople, but neither pipeline reliably produces candidates who excel at all three, which is why specialized recruiting knowledge matters.
FDEs earn a premium over standard software engineers, with median total compensation at $173,816 and top-tier roles at companies like Palantir reaching $238,000 to $486,000. The client-facing dimension and ownership of customer outcomes commands its own premium that the market has already priced in.
Beyond core technical skills like Python, TypeScript, SQL, and cloud infrastructure (AWS, GCP, Kubernetes), FDEs need customer empathy, clear communication under pressure, and product sense to decide what to build and why. For AI-focused roles, add agentic orchestration, model evaluation frameworks, and AI observability tooling to the requirements.
AI startups adopted the model Palantir pioneered, reshaping it for compressed cycles where one engineer owns an entire customer relationship from first call to production deployment. The role solves a fundamental gap: great software alone isn't enough when customers need someone to close the distance between what a product can do and what they need it to do in their specific environment.
Join world-class companies that build their teams with Paraform.
