In early 2026, Anthropic posted a wave of Forward Deployed Engineer roles paying $300K–$480K total comp. OpenAI followed with similar listings. Sierra, the enterprise AI startup built by ex-Salesforce leadership, had been running this model for longer. The job postings all describe something that looks less like a traditional software engineering role and more like a highly technical account manager who ships code — and the compensation reflects how hard that combination is to find. At AgenticCareers.co, we track these postings daily and have spent time understanding what the role actually requires in practice. Here is what we have learned.
The Forward Deployed Engineer title has roots at Palantir, which invented it in the mid-2000s to solve a specific problem: government and enterprise customers buying complex data platforms needed someone who could sit on-site, understand the customer's actual data and workflows, and build the integrations that made the software work in their environment. Palantir's insight was that the gap between "software sold" and "software working" required a distinct engineering archetype — not a sales engineer who demos, not a consultant who recommends, but a software engineer who codes inside the customer's systems and owns the outcome. That model worked, and the title spread.
The resurgence of FDEs at frontier AI labs follows the same structural logic. A frontier model is extraordinarily capable in the abstract, but deploying it reliably inside a Fortune 500 customer's environment — against their proprietary data, under their security constraints, integrated with their existing tooling — requires a kind of engineering that research teams and product engineers do not naturally focus on. FDEs are the people who close that gap, and they are being hired at a pace that suggests the market has concluded this role is not a luxury.
What FDEs Actually Do
- Customer onboarding and scoping — FDEs begin every engagement by understanding the customer's existing systems, data architecture, and deployment constraints. This is not a sales conversation — it is a technical audit that informs everything the FDE will build over the engagement. The quality of this scoping work directly determines whether the deployment succeeds.
- Integration engineering — Most enterprise AI deployments require connecting the vendor's platform to the customer's existing infrastructure: internal APIs, databases, document stores, identity providers, and SaaS tools. FDEs write the integration code — connectors, adapters, data pipelines — that makes the AI system a functional part of the customer's environment rather than an isolated prototype.
- Production deployment — FDEs are responsible for getting the system from demo to production. This means managing deployment infrastructure, handling the security and compliance requirements of the customer's environment (SOC2, HIPAA, FedRAMP where relevant), and ensuring the system performs reliably at scale — not just in a controlled evaluation.
- Evals on customer data — A frontier model evaluated on public benchmarks behaves differently on a customer's actual data. FDEs design and run evaluation harnesses against the customer's real inputs, measure performance on the tasks that actually matter to the customer, and iterate on prompts, fine-tuning, or system architecture based on what the evals reveal. This is one of the most important and underappreciated parts of the role.
- Incident response — When a deployed AI system produces bad outputs in a live customer environment, the FDE is the person who diagnoses and fixes it. This requires debugging skills adapted to probabilistic systems — the failure may not reproduce, the root cause may be a distribution shift in the customer's data, and the fix may need to happen without disrupting production. Being on the hook for live deployments is a defining characteristic of the FDE role.
- Training customer teams — FDEs spend meaningful time transferring knowledge to the customer's own engineers and operators: how to maintain the integrations, how to interpret eval results, how to extend the deployment as needs evolve. The goal is not permanent dependency on the vendor — it is making the customer self-sufficient enough that the deployment survives the FDE's eventual rotation to a new engagement.
Skills and Tools
The skill profile that makes someone effective as an FDE is distinctive and not easy to develop casually. It requires strong software engineering fundamentals combined with customer-facing communication skills — a combination that is genuinely rare and almost always assembled deliberately across multiple roles.
- Senior software engineering fundamentals — Strong enough to own production systems without a senior engineer supervising the work. Understanding of API design, async patterns, deployment infrastructure, and operational discipline. FDEs are not junior engineers with good people skills; they are expected to make sound architectural decisions independently in the customer's environment.
- LLM APIs and agentic frameworks — Hands-on fluency with the Anthropic, OpenAI, and Google APIs. Understanding of context windows, tool use, structured outputs, and the behavioral differences between model families. Working knowledge of agentic frameworks for customers who need multi-step systems.
- Enterprise integration patterns — SAML, OIDC, SSO, and the identity and access management patterns that enterprise customers require. Familiarity with SOC2 workflows and the operational requirements of security-conscious deployments. The ability to navigate enterprise procurement and IT constraints while still shipping code.
- Evaluation tooling — Experience building eval harnesses — whether with platforms like Braintrust and Langfuse or custom implementations — and the judgment to design evals that actually measure what matters in a customer's use case rather than vanity metrics.
- Customer-facing communication — The ability to translate technical findings into business language for executives and technical constraints back to the team. FDEs regularly operate in meetings with non-technical stakeholders, and the clarity of their communication directly affects whether deployments succeed.
- Observability — Ability to instrument AI systems for latency monitoring, cost tracking, output quality measurement, and anomaly detection. Knowing what to watch in a live deployment and how to act quickly when something drifts.
Salary Range (2026)
Compensation for FDEs reflects the scarcity of the skill combination and the direct revenue impact these engineers have on enterprise deals. Based on AgenticCareers listings in early 2026:
- Frontier labs (Anthropic, OpenAI, Google) — Total compensation ranges from approximately $300K at entry level to $480K and above at senior level in the US. Equity grants at this tier are substantial and refreshed annually for strong performers.
- Palantir — The originator of the role pays $240K–$420K total comp depending on seniority and region. Palantir's FDE culture is the most established in the industry, and the training ground reputation carries weight in the market.
- Enterprise AI firms (Sierra, Harvey, Hebbia, Glean) — Total comp ranges from $220K to $400K depending on company stage, seniority, and equity percentage. Earlier-stage companies compensate with higher equity stakes; later-stage firms offer more cash certainty.
- Travel requirements vary widely — Many FDE roles require regular on-site time with customers. Always read the JD carefully — "up to 50% travel" means something different than "based at customer site." Remote FDE roles exist but are less common than remote engineering roles at the same comp level.
FDE vs Solutions Engineer vs Applied AI Engineer
The FDE title is frequently confused with two adjacent roles, and the distinction matters if you are considering this career path. Solutions Engineers are pre-sales: they run technical demonstrations, answer architectural questions during the sales process, and help prospects understand whether the product fits their needs. They may write code, but they rarely own production deployments. The handoff to a customer success or professional services team happens at contract signing. FDEs, by contrast, own the deployment post-sale — they write production code and are accountable for whether the system works in the customer's environment.
The comparison with Applied AI Engineers is more subtle. Applied AI Engineers at frontier labs are typically building and improving the vendor's own products and deployment infrastructure — they work on the platform that FDEs then deploy for customers. Applied AI Engineers may be customer-facing in some contexts (especially at enterprise AI firms), but their primary output is the product itself, not a customer-specific deployment. FDEs are applied specifically to a customer's environment, working with whatever the product team has shipped.
In practice, the line between these titles blurs at smaller companies. At Palantir, the FDE role has decades of meaning something specific. At a seed-stage AI startup, "Forward Deployed Engineer" might mean "the person who does implementation work promised in sales calls." Always read the actual job description carefully.
How to Become an FDE
Senior Software Engineer at an enterprise-facing company
This is the most common path. Senior engineers who have experience deploying systems inside large organizations — navigating procurement, IT security reviews, and enterprise integration requirements — have already developed many of the instincts the FDE role requires. The investment is in developing customer-facing communication skills and LLM API depth. Most engineers making this transition find the customer-interaction component the harder adjustment; the technical skills translate more directly than expected.
Consultant or Solutions Engineer with strong coding skills
People who have spent time in consulting or solutions engineering roles have the customer-facing muscle that many engineers lack. The gap is typically in depth and ownership of production engineering — moving from recommending architectures to being accountable for them running in production. The transition requires deliberately taking on more implementation work and building fluency with LLM APIs and evaluation tooling.
Palantir FDE moving to a frontier lab
Palantir's FDE program is the longest-running at scale, and alumni are in high demand at Anthropic, OpenAI, and enterprise AI firms. Having trained inside the most rigorous customer-embedded engineering culture in the industry is a meaningful credential. The transition involves developing depth on frontier LLM capabilities — Palantir's stack is distinct from the frontier lab ecosystem — but the core operating model transfers directly and commands a premium in the market.
What to Ask in the Interview
FDE roles vary significantly in what they actually involve. These questions surface misaligned expectations before you accept an offer:
- What percentage of time do FDEs spend on-site with customers? Travel expectations range from minimal to near-constant, and the answer reveals whether the role is truly embedded or primarily remote with occasional visits.
- Does this role report into customer success or engineering? Reporting line signals how the company thinks about FDEs. Customer-success reporting means more account management weight; engineering reporting means more production ownership. Know which you are signing up for.
- What is the ratio of coding time to meetings in a typical week? Some FDE roles are primarily technical implementation; others involve significant coordination overhead. Ask explicitly how much of the role is actually writing code.
- What does handoff look like at the end of an engagement? This reveals whether the company has a mature transition model or whether FDEs remain indefinitely responsible for systems they built months ago.
Related reading
For more context on adjacent high-leverage roles: What is an Applied AI Engineer? covers the role that often sits on the product side of the same customer deployments FDEs handle, and the AI Agent Engineer salary guide 2026 provides detailed compensation data with breakdowns by company stage, geography, and seniority that applies to FDE roles as well.