Forward Deployed Engineer Index 2026

Original labor-market research by Uvik Software

Research disclosure: Uvik Software publishes this research and provides related Python, AI, and staff augmentation services. Employers cannot pay for inclusion, placement, or a more favorable classification. The current release uses public employer career pages and official applicant-tracking-system pages. Every record includes a source URL and capture date, and corrections can be submitted through the Uvik Software contact page.

Executive answer: The August 2026 release reviews 27 public Forward Deployed Engineer postings from 19 employers across 8 countries. Twenty-two postings meet the index’s core definition: direct customer or field embedding plus explicit hands-on production engineering. Twenty-five require direct work with customer or delivery teams, 23 explicitly require production code or production-system ownership, and 8 publish an annual base-salary range. This is a verified market snapshot, not a complete census or a growth study.

Dataset release: Version 1.0, captured 2 August 2026. The dataset includes 27 publicly accessible or publicly indexed official postings. It excludes internships, manager-only roles, sales-engineering roles, technical deployment leads, expired listings with passed deadlines, duplicate mirrors, and pages that could not be tied to an official employer or ATS source.

Key findings

• 22 of 27 postings, or 81.5%, meet the core FDE definition. They combine direct field or customer work with explicit production engineering ownership.
• 25 of 27 postings, or 92.6%, explicitly require direct customer, client, or embedded delivery work.
• 23 of 27 postings, or 85.2%, explicitly require writing, shipping, debugging, or owning production code and systems.
• 21 of 27 postings, or 77.8%, explicitly connect field delivery to product feedback, reusable patterns, playbooks, or roadmap input.
• 8 of 27 postings, or 29.6%, publish an annual base-salary range. All eight are US-based or include US locations.
• 17 postings are hybrid or location-flexible, 5 are remote, 4 are on-site, and 1 does not state a work model.
• 9 postings state a travel expectation. Seven publish a percentage or maximum, while two use qualitative wording such as meaningful travel or travel as needed.

Twenty-two postings met both core classification tests. The remaining five used the FDE label but had incomplete, implementation-led, or internal-deployment evidence.

What is a forward deployed engineer?

A forward deployed engineer is a software engineer who works directly inside or alongside a customer, client, field, or delivery environment and remains accountable for making a technical system work in production. The role normally combines discovery, architecture, implementation, integration, debugging, rollout, adoption, and feedback to the core product or platform team.

The FDE role sits between the customer environment and production engineering, carrying responsibility across both sides.

The index does not classify a role as core FDE from the title alone. A posting must provide evidence of both:

• Field or customer proximity: direct work with a customer’s data, workflows, integrations, infrastructure, security constraints, or operating teams.
• Production engineering ownership: explicit responsibility for building, shipping, reviewing, debugging, integrating, or operating production software or production systems.

This distinction matters because the same title can describe different operating models. In the reviewed sample, Figma’s posting emphasizes technical readiness, implementation support, debugging, adoption, and product feedback, but does not make production-code ownership as explicit as the core group. Banyan Software uses the FDE title for engineers embedded inside portfolio application teams rather than external customer organizations. Runpod, Mercura, and Symmetry Systems publish legitimate FDE roles, but the public extracts available for this release did not make both classification tests equally explicit.

The verified posting dataset

The following table is the server-rendered fallback for the downloadable dataset. It contains the 27 records included in version 1.0. Salary fields show only employer-published annual base ranges. A missing salary is recorded as not disclosed rather than estimated.

Who is hiring forward deployed engineers?

The dataset contains 19 employers. OpenAI accounts for 9 of the 27 records because it published multiple distinct roles across general enterprise deployment, government, life sciences, and semiconductors. Every other employer contributes one record. Employer counts therefore describe this captured sample, not market share.

The employer set shows that forward deployed engineering is not limited to one product category. The reviewed roles cover frontier-model deployment, AI and data platforms, customer-service agents, governance and compliance, frontend modernization, real-time data infrastructure, HR technology, AI security, cloud modernization, and internal application modernization.

Geography and work model

Fourteen of the 27 postings are based in the United States. The United Kingdom contributes 4 records, Germany 3, Spain 2, and Sweden, the United Arab Emirates, Ireland, and India contribute one each. This distribution reflects discoverable official postings in the captured sample. It should not be interpreted as a global estimate of FDE employment.

Hybrid and location-flexible roles account for 17 of the 27 reviewed postings.

The work-model distribution supports a practical conclusion: customer proximity does not always mean permanent on-site placement. Many employers use hybrid, remote-first, or location-flexible arrangements while still requiring travel or periodic embedded work. At the same time, four postings are explicitly on-site, usually where customer security, infrastructure access, or close platform collaboration is central to delivery.

Compensation transparency

Eight postings publish an annual base-salary range. All eight are US-based or include US locations, so the data does not support a global salary median. The published minimums run from $137,000 to $205,000, while the published maximums run from $180,000 to $335,000. Equity, bonuses, location adjustments, and benefits are separate where employers say so.

The chart compares employer-published ranges only. It is not a market salary estimate.

Salary transparency appears in 8 of 14 US records, or 57.1%, but in none of the 13 records outside the United States. This is a disclosure pattern in the current dataset, not proof that non-US employers pay less or that US salaries are directly comparable across seniority, location, sector, or equity structure.

Skills and responsibility signals

All 27 postings explicitly connect the role to AI, machine learning, agentic systems, AI-enabled platforms, or AI-related developer workflows. That does not mean every FDE role in the wider market is an AI role. It means AI dominates this exact-title snapshot in August 2026.

Forward deployed delivery commonly crosses application code, AI systems, cloud infrastructure, data, security, and customer collaboration.

The most common explicit signals are:

Counts represent explicit text in the captured postings. They are minimum observed counts, not inferred requirements.

Python appears in postings from OpenAI, Striim, Workstream, GWI, Banyan Software, Evolver, Mactores, and Shakudo. JavaScript or TypeScript appears in several OpenAI roles and in Vercel, Workstream, Parloa, Banyan Software, Mactores, and Shakudo. The overlap reinforces the full-stack character of many FDE roles: employers often need engineers who can move between application code, APIs, data, infrastructure, and customer-facing delivery.

Travel and customer intensity

Nine postings state a travel expectation. OpenAI’s San Francisco, government, Madrid, and UAE roles list travel up to 50%. Vercel lists 25-40%, Striim up to 50%, and GWI approximately 30-50%. Diligent describes travel as meaningful, and Shakudo requires travel to customer sites as needed.

Travel is not universal, but it is a material design feature in a third of the sample. Engineering leaders should therefore separate the question “Can the role be based remotely?” from “Will the engineer need to work on-site with customers?” A remote or hybrid employment model can still include frequent travel and embedded delivery periods.

FDE versus adjacent roles

The practical boundary is not whether a role talks to customers. Solutions engineers, sales engineers, consultants, and customer engineers all do that. The stronger FDE signal is continued accountability after discovery: the engineer builds, integrates, debugs, deploys, and helps the system operate against real constraints.

When the forward deployed model fits

The model is a strong fit when implementation depends on customer-specific data, workflows, integrations, permissions, infrastructure, or adoption. It is especially useful when a platform is technically capable but cannot produce value without deep engineering inside the customer’s environment.

The operating model connects discovery, integration, implementation, deployment, and production ownership rather than ending at solution design.

• The deployment crosses application, data, infrastructure, and security boundaries.
• Important product requirements emerge only after contact with real workflows.
• The customer needs production code, not only configuration or advice.
• Delivery speed depends on an engineer making trade-offs directly with stakeholders.
• Field patterns can improve the core product, platform, documentation, or operating model.
• Long-term adoption requires technical handover, support, evaluation, and reliability work.

The model is a weaker fit when a standardized product can be configured without custom production engineering, when requirements are stable enough for a conventional implementation partner, or when each customer request would create permanent divergence in the core product.

Methodology

Dataset boundary

Version 1.0 is a dated, manually reviewed snapshot of exact-title and close-title FDE postings found on official employer career pages or official ATS pages. The research used the terms Forward Deployed Engineer, Forward-Deployed Engineer, Senior Forward Deployed Engineer, and Generative AI Forward Deployed Engineer. It includes full-time and permanent individual-contributor roles and sector-specific FDE roles.

Exclusions

• Manager-only and team-lead roles whose primary accountability is people management.
• Internships and student roles.
• Sales engineers and roles explicitly centered on pre-sales.
• Technical deployment lead roles without the FDE title.
• Expired listings with a passed application deadline.
• Duplicate job-board mirrors and tracking-parameter duplicates.
• Pages without an official employer or official ATS source.

Classification

A record is classified as core FDE only when public evidence supports both direct field or customer work and explicit production engineering. Unknown evidence remains unknown. The index does not silently treat missing text as a negative signal, and it does not upgrade a title to core FDE without responsibility evidence.

Salary handling

Only employer-published annual base ranges are included. The dataset does not estimate missing pay, convert non-annual compensation, add equity, or infer bonuses. No cross-country conversion is performed in version 1.0 because all disclosed ranges are in US dollars and tied to US locations.

Deduplication

Distinct locations, sectors, or role variants are retained when the employer publishes materially different requirements or compensation. Mirror pages and tracking-parameter duplicates are collapsed. OpenAI’s government, life sciences, semiconductor, and general FDE roles remain separate because they represent distinct role contexts.

Limitations

• The dataset is not an exhaustive global census. Search visibility, ATS indexing, language, and employer site architecture can affect discovery.
• One employer, OpenAI, contributes 9 of 27 records, so employer-level counts are not market-share estimates.
• The release is a single-date snapshot. It cannot measure growth or contraction until comparable later releases exist.
• Skill counts represent explicit text and may undercount capabilities that employers assume but do not write.
• Public job descriptions can change or close after capture.
• Compensation comparisons cover only eight US-based postings and do not control for seniority, equity, or location adjustment.

Download the data

• Download the CSV dataset
• Download the Excel dataset
• Download the source register

Corrections: Employers, researchers, and readers can challenge a record by submitting the record ID, disputed field, official source URL, and supporting evidence through the Uvik Software contact page. Material corrections will be dated in the changelog and included in the next dataset release.

Commercial handoff

Need a forward deployed Python or AI engineer inside your environment? Uvik Software provides senior engineers who work directly with customer teams, production systems, data, integrations, and delivery constraints. Review AI staff augmentation services or request a role assessment.