The 800% FDE Spike: Why LinkedIn Title Search Is the Wrong Tool
Forward Deployed Engineer postings jumped 800% in 2025. Here's how to source FDEs by hunting dual signals instead of chasing a LinkedIn title.
Every founder building anything customer-facing in AI suddenly wants a Forward Deployed Engineer. The postings are real, the comp is real, and the talent pool is not where the title search says it is. If you're typing "Forward Deployed Engineer" into LinkedIn Recruiter and wondering why every result looks like a Palantir lifer, you're sourcing against a label that the market only just adopted.
The spike is real, the title is a trailing indicator
The Financial Times clocked FDE job postings up more than 800% from January through September 2025. Lightcast counted 922 listings by November, a 5x jump year-over-year. BCG renamed its BCGX engineering and design staff to Forward Deployed Engineers on May 1, 2026. EY launched FDE roles across the UK and Ireland in April 2026. Salesforce publicly committed to a 1,000-person FDE org. OpenAI, Anthropic, Databricks, and Cohere are all building.
Here's the trap. The title only became standard in mid-2025. Almost every great FDE working today, the people you actually want, has spent most of their career labeled something else: Solutions Architect, Sales Engineer, Implementation Engineer, Technical Delivery Engineer, or, if they're Palantir, "Forward Deployed Software Engineer" or "Delta." Searching LinkedIn for the literal phrase "Forward Deployed Engineer" misses the entire pre-2025 talent base, which is the only base with real receipts.
This is why FDE recruiting fails the standard keyword playbook. The signal you want, ships production code AND survives an executive readout, is not encoded in a job title. It's encoded in a pattern of behavior across GitHub, a resume, and a customer-facing track record.
What an FDE actually is, and why it breaks keyword search
The clearest definition comes from Aaron Zelinger, ex-Palantir FDE and now CEO of Closure: FDEs are a technical bridge that eliminates the bureaucratic layer of sales and account managers, doing customer discovery without a non-technical filter. They sit with a customer, find the workflow worth automating, and then go write the code that automates it. They ship.
That dual mandate is why adjacent titles overlap so heavily. A Solutions Architect at a cloud vendor sells a reference architecture but doesn't merge to main. A Sales Engineer demos and disappears. An Implementation Engineer configures but rarely owns net-new code. An FDE does all of it. The differentiator, per Pragmatic Engineer's analysis, is that FDEs work with customers AND contribute to the product.
The technical floor is unromantic: Python or TypeScript, SQL and data systems, cloud infrastructure. The AI-focused version adds LLM deployment and agentic workflow experience. The shape is T-shaped: broad enough to own a deployment end to end, deep enough to debug a production incident at 2am with the customer's CISO on the line.
If you screen on AI keywords (LLM, RAG, agents, vector DB) you'll get a flood of resume optimizers. The actual signal is more boring and more useful: did this person ship something that real users touched, and can they explain it to a non-engineer without condescending.
The dual signal: production code plus customer receipts
You're hunting two things at once, and most sourcing tools are built to hunt one.
Signal one: production code that shipped. GitHub matters here, but not in the "1,000-star side project" way. Look for repos with deployment configs, dockerfiles, infrastructure code, error handling, and commits that span more than three weeks. Look for founders who shipped a startup, even a failed one. Look for engineer #2 or #3 hires at YC companies who had to wear five hats.
Signal two: customer-facing receipts. This is the part that breaks GitHub-only tools. Public writing that translates a hard technical problem to a non-technical reader. Conference talks. A "field CTO" or "deployment strategist" stint. A Solutions Architect role at a company whose customers are themselves technical (Snowflake, Databricks, Vercel, Modal). Time at a consulting firm where the engineer actually wrote code, not slides.
Most candidates with both signals don't show up in a LinkedIn title search and don't show up in a GitHub-only search. They show up when you query both surfaces at once with a description of the human, not a string match on a job title. This is the specific friction we built Refolk for: you describe the FDE you want in plain English, including the dual signal, and you get a ranked shortlist drawn from GitHub, LinkedIn, and the open web in the same pass.
Where the FDEs actually are right now
A few seed pools beat any title search.
The Palantir "Delta" diaspora
No company has shaped the role more, and until 2016 Palantir had more FDEs than software engineers. When Foundry launched, many rotated back into core product, but the alumni pool that left has seeded almost every interesting AI company. Lightcast confirms Palantir was the top company posting FDE listings between 2010 and 2025, followed by PwC and Salesforce.
The Palantir Delta diaspora is the highest-yield single seed for FDE sourcing in 2026, full stop.