LinkedIn Says FDE Demand Is Up 226%. The US Pool Is 3,486.

On September 29, 2026, at Talent Connect, LinkedIn's Dan Reid and Prashanthi Padmanabhan dropped a number into the Hiring Assistant announcement that most sourcers skimmed past: demand for forward-deployed experts is up 226% in the last year, the single fastest-growing technical title on the platform. AI forensic analyst roles are up 169%. AI safety is up 91%. Those are demand signals from the buy side, not supply signals from the people side, and the gap between them is the whole story.

If you want to actually fill one of these reqs, the 226% is a trap. The pond is tiny, the pond is concentrated, and almost everyone fishing it is using the wrong keyword.

How many Forward Deployed Engineers actually exist

In Refolk's index of professional profiles, 3,486 people in the US currently carry a Forward Deployed Engineer, FDE, or FDSE title. That is the ceiling, not the floor, and it is already being drained.

A mid-2026 market scan counted 224 open FDE positions across 39 AI companies. Palantir alone had 76 open FDE-family reqs as of September 2, 2026, with 49 in the US. Do the arithmetic:

• 3,486 US FDE-titled profiles in Refolk's index
• ~224 open FDE reqs across the top 39 AI employers
• Roughly 1 in 15 US FDEs is being actively pipelined right now

That is before you subtract the FDEs who just started, the ones vested into golden handcuffs at Palantir, and the ones who have already left the function. The real addressable pool for any given search is a few hundred people.

3,486

The supply map, by segment

The US-to-rest-of-the-named-world ratio of roughly 2.94x is why so many European AI labs quietly staff their deployment function with Americans on secondment. The talent is not there in the volume the headline implies.

Why the 226% is a demand signal, not a talent signal

The 226% growth number measures reqs posted and search activity, not humans who can do the job. The humans grew at roughly the rate humans grow, which is not very fast. That is the whole mechanism behind the salary explosion.

Forward Deployed Engineer, for the one person left who still needs the definition: an engineer who embeds inside a single customer, owns the implementation of a platform end to end, and stays until it runs in production. Shyam Sankar, Palantir's CTO, coined the term in 2007 after Alex Karp made an offhand comment that in a French restaurant the wait staff are really part of the kitchen. Palantir's internal name for the role is still Delta, where a Delta sits inside Business Development and deploys the platform to one customer, and a Dev sits inside Product Development and builds Foundry and Gotham for everyone.

That Delta vocabulary matters for sourcing. If you only Boolean on "Forward Deployed Engineer" on LinkedIn, you are fishing a pond of under four thousand US profiles. Boolean on "Delta" AND "Palantir" and you surface the actual alumni cohort, including the ones who moved on to titles like Deployment Strategist, Solutions Architect, Applied AI Engineer, or just plain Engineer at a startup where they are the fourth hire. This is the exact gap Refolk closes: you describe the person in plain English ("ex-Palantir Delta now working on applied AI at a Series B, based in New York") and get a ranked shortlist across GitHub, LinkedIn, and the open web without hand-tuning a Boolean.

Where real FDEs actually work

Palantir is still the center of gravity, but the named-employer distribution has splintered fast. In Refolk's index, the current US FDE employers that show up most often outside Palantir are:

• The OpenAI Deployment Company, the spun-out entity that proves OpenAI has operationalized the function as a separate legal unit
• Sixfold, insurance AI, NY-weighted
• Gecko Robotics, industrial inspection
• Amp, logistics and recycling automation
• Roboflow, computer vision
• Cresta, contact-center AI
• Ramp, which stood up an FDE function under ex-Palantir leadership
• Deloitte, which is now billing "Palantir FDEs" as a consulting line (its current req closes 10/30/2026)

The Illinois Tech career page captures the current buyer list cleanly: FDEs are in demand at OpenAI, Anthropic, Ramp, Salesforce, industrial and healthcare AI startups, and, after a $1 billion investment announced in 2026, Amazon Web Services.

If you only Boolean on the exact title, you are fishing a pond of fewer than four thousand people.

The one career arc worth memorizing

Kevin Bai scaled Palantir's FDE function, left to found the FDE org at Rippling, and now works on Anthropic's Applied AI team. That arc, Palantir to high-growth non-AI company to frontier lab, is the modal path for the most senior people in the market. If a candidate's history rhymes with Bai's, pay the premium. If their history is Palantir to consulting to Deloitte FDE, you are looking at a different job entirely.

The Palantir alumni problem, and the Encord exception

Every FDE buyer is competing for the same roughly five-year Palantir alumni cohort, which is why frontier labs now pay a 60 to 150% premium over Palantir base. If you are sourcing against this market and you only know how to find Palantir alumni, you will lose every one of them to OpenAI's comp band.

Comp bands worth repeating in any pitch:

• Palantir: median base around $167,000, total comp around $215,000
• OpenAI mid-level: base $160,000 to $280,000, total comp $350,000 to $550,000 with equity and bonus
• Frontier labs overall: 60 to 150% premium over Palantir

Europe is worse. In the UK, Germany, and Canada combined, Refolk's index shows 1,187 FDE and Deployment Strategist profiles, and Palantir supplies roughly 28% of the named employers, versus 16% in the US. London is the single densest non-US FDE city on earth. Berlin is a distant second. The only non-Palantir London employer with multiple FDEs in the sample is Encord, with 3, which quietly makes it the single most interesting poaching target for anyone running Palantir FDE recruiting out of a European office.

What LinkedIn's "AI forensic analyst" line actually means

The +169% AI forensic analyst number in the same LinkedIn release is a label collision, not a role. It bundles trust and safety engineers, model-eval specialists, and incident-response people who almost never carry the exact "AI forensic analyst" title.

If you take the keyword literally and search LinkedIn for it, you will find a few hundred profiles, most of them repurposed e-discovery and digital-forensics people from the compliance world. The actual humans driving the demand sit in Trust and Safety orgs at OpenAI, Anthropic, and Scale AI, under titles like:

• Policy Engineer
• Red Team Lead or Red Teamer
• Model Evaluations Engineer
• Safety Researcher
• Trust and Safety Engineer

AI forensic analyst hiring is really trust and safety hiring with a new coat of paint. If your intake call starts with "we need an AI forensic analyst," your first job is to translate the req into one of the five titles above, then source against the orgs that actually use them.

A sourcing playbook that beats the exact-title string

The job is to find the humans who can do FDE work, not the humans who happen to have "FDE" in their current title. Six Boolean plays that outperform the exact-title string:

1. "Delta" AND "Palantir" on LinkedIn and GitHub bios. Finds the alumni who left under a different title.
2. "Deployment Strategist" as a title, Palantir and OpenAI only. The second-most reliable FDE proxy.
3. "Forward Deployed" NOT "Palantir" to surface the people who adopted the vocabulary at a new employer, a strong signal of real identity attachment to the role.
4. GitHub commits to customer-specific repos at companies known to run FDE models (Palantir, Sixfold, Gecko). The commit graph reveals who actually ships to one customer vs. who ships to a product.
5. "FDSE" as a legacy Palantir internal acronym. Mostly surfaces 2017 to 2022 Palantir cohorts, which is the sweet spot for frontier lab buyers.
6. "Applied AI Engineer" at OpenAI, Anthropic, and Cognition. In 2026 this is the de facto FDE title at labs that refuse to use the Palantir vocabulary.

The reason the exact-string approach keeps failing is that FDE is an identity, not a job code. People who did the work at Palantir recognize each other across any title. The sourcing tooling has to recognize them too, which is a different problem than keyword matching.

The Deloitte problem: filter on equity and embed, not title

A Deloitte FDE and an OpenAI FDE share a title and almost no job overlap. Filter your pipeline on equity grant and direct customer embed, not on the three letters.

Deloitte is now hiring "Palantir FDEs" as a billable consulting line, with a current req closing 10/30/2026. Technically those are FDEs. In practice, a Deloitte FDE is a consultant on a Palantir project with Deloitte's utilization targets, Deloitte's bench, and Deloitte's bonus structure. The lab version of the role is equity-heavy, single-customer, production-owning, and does not have a utilization target.

If you are a founder running your own search, the three filters that separate the two populations:

• Equity package in the top quartile of the band (lab FDEs get frontier-lab equity, consulting FDEs get consulting bonuses)
• Named single customer in the current scope (not a portfolio)
• Production ownership past deployment (not a report-and-handoff model)

The 70% chemical-usage reduction OpenAI's FDE team delivered at John Deere, the most-cited customer anecdote in the whole category, happened because the OpenAI engineers stayed in the Deere environment past deployment. A consultant leaves at go-live. An FDE does not.

The early-career angle nobody is working

Palantir's new-grad FDE program is still the only true early-career FDE funnel in the market. Every other buyer is fishing the same five-year Palantir alumni pool, which is why they are losing on comp.

If you are a founder who cannot pay OpenAI comp, the arbitrage is to recruit Palantir new-grads at the 18-to-30 month mark, before they have fully vested but after they have run at least one deployment solo. That cohort is in Refolk's index with reasonable fidelity because Palantir new-grads tend to keep their title transitions clean on LinkedIn (Intern to FDE Associate to FDE). Pulling the cohort by graduation year and Palantir start date, then filtering on current title and tenure, is the usual way to catch the ones who are quietly open. Describe that pattern in plain English to Refolk and the shortlist comes back ranked across LinkedIn, GitHub, and the open web.

FAQ

What is the difference between a Forward Deployed Engineer and a Solutions Engineer?

A Solutions Engineer sells and prototypes. A Forward Deployed Engineer owns implementation and production delivery for a single named customer, usually with equity exposure to the outcome. The Solutions Engineer leaves after the deal closes; the FDE stays until the system runs in production. In practice, Solutions Engineers at platform companies sometimes transition into FDE roles at frontier labs, which makes them a legitimate secondary feeder pool when the Palantir alumni cohort is tapped out.

Is "AI forensic analyst" a real job title I can source against?

Not usefully. LinkedIn's +169% number for AI forensic analyst roles bundles trust and safety engineers, model-eval specialists, and incident response people who almost never carry that exact title. Source against Trust and Safety orgs at OpenAI, Anthropic, and Scale AI, and translate the req into Policy Engineer, Red Team Lead, Model Evaluations Engineer, or Safety Researcher before you start keyword-matching. The exact phrase will return mostly repurposed e-discovery people from compliance firms.

How do I find ex-Palantir FDEs who have changed their title?

The highest-yield Boolean is "Delta" AND "Palantir", because Delta is the internal Palantir term that FDE alumni still use in bios and GitHub profiles even after they drop the formal title. "FDSE" and "Deployment Strategist" are the next two highest-signal strings. Beyond keywords, work the GitHub commit graph for people whose history shows single-customer repo patterns, and cross-reference with Palantir's new-grad class years to surface the two-to-five-year alumni cohort that frontier labs are actively pricing.

Where should I look for FDEs outside the US?

London first, by a wide margin. In Refolk's index, the UK, Germany, and Canada together hold 1,187 FDE and Deployment Strategist profiles, versus 3,486 in the US, and London supplies the plurality of that European pool. Palantir dominates even more heavily in Europe (about 28% of named employers vs. 16% in the US), and the single non-obvious London poaching target is Encord, the only non-Palantir London employer with multiple FDEs in the sample.