Forward Deployed Engineer (FDE)

What does this guide cover?

This page is the overview. Each question that deserves more than a section has its own page, and the skills the job tests are taught one at a time in The Forward Deployed Playbook.

• Companies hiring Forward Deployed Engineers Who posts the role, how many openings, where, and what each company means by the title.
• How to become a Forward Deployed Engineer The skills the job actually tests, and the 7-episode playbook that teaches them.

What does a Forward Deployed Engineer do day to day?

An FDE spends the week between two codebases: the vendor’s platform and the customer’s estate. The deliverable is running software that solves one specific customer problem, which means the job is as much finding out what the problem actually is as it is writing the code. A typical engagement moves through four kinds of work:

1. Discovery. Sitting with the people who will use the system, reading the process they follow today, and finding the exceptions nobody wrote down.
2. Integration. Connecting the model or platform to the customer’s data, identity, permissions and legacy systems, then getting it through their security and architecture reviews.
3. Delivery. Shipping a first release that real users rely on, measuring whether it moved the business number it was bought for, and iterating.
4. Handover. Leaving behind monitoring, runbooks and a customer team that can operate and extend the system after the FDE moves on.

The mechanism that makes the role exist: a frontier model or a data platform does nothing for a customer until someone encodes that customer’s business into it. Sales can promise that work and documentation can describe it, but somebody has to do it on site.

Where did the Forward Deployed Engineer come from?

Palantir. The company built its entire delivery model around engineers placed inside the customer organisation, working on the customer’s problem on top of Palantir’s platforms rather than shipping a product and handing over documentation. Palantir pairs the FDE with a Deployment Strategist, who owns the customer relationship and the problem definition while the engineer owns the build. That is a decade-plus of precedent, and it is the single most useful fact for understanding why the AI labs picked this title instead of inventing one. They were not naming a new job. They were importing a delivery model that already had a track record in a business with the same shape: expensive platform, high-context customer problem, and no clean handoff.

Our Palantir salary guide covers the original FDE track in detail, including how the role sits in that company’s career structure.

Which companies are hiring Forward Deployed Engineers?

Rung’s November 2026 census counts 394 open FDE roles across the 13 companies it tracks. Four employers hold most of them:

Anthropic posts Forward Deployed Engineer, Applied AI, Meta hires Forward Deployed Engineer, Enterprise AI, and ByteDance’s BytePlus has posted the role in Singapore with a description that is entirely about building AI agents. The full breakdown by company, location and remote share is on companies hiring Forward Deployed Engineers.

How much does a Forward Deployed Engineer earn?

Few FDE postings publish pay, so every number here comes from a thin sample. Rung’s census found 32 US postings with a disclosed band. The median band was $185K to $300K base, the floor was around $132K, and the top was $445K plus equity. Frontier labs add equity that can match or exceed base, while Palantir and most non-US postings disclose nothing. The Palantir salary guide has the longest-running data on the original track.

What skills does a Forward Deployed Engineer need?

Postings ask for a short, consistent list of hard skills. Of the 306 descriptions Rung could read, 77% ask for Python, 64% for APIs and integrations, 63% for a major cloud, and 99% stress customer-facing work. Only 33% name LLMs explicitly, which tells you the labs are hiring for delivery judgment first and model knowledge second.

The judgment is the hard part, and it is what the playbook teaches. Each episode takes one skill that decides whether a deployment survives contact with the customer:

1. EP01 · Measure business outcomes Measure the outcome, not the sprint Why story points say nothing about whether a deployment worked, and what to measure instead.
2. EP02 · Turn business context into system behavior Context engineering: turning business rules into agent behavior How domain knowledge, rules and permissions become the context an agent runs on.
3. EP03 · Product management fundamentals From AI pilot to production Scoping a first release that survives contact with real users. Coming soon.
4. EP04 · Build with AI Building with coding agents without trusting them blindly Where agents speed delivery up, and the judgment you still owe the code. Coming soon.
5. EP05 · Build across enterprise constraints Enterprise AI integration Legacy systems, security review and debugging across the whole stack. Coming soon.
6. EP06 · Engineer for trust and sustained operation Human in the loop, audit trails and operating AI What it takes for a customer team to trust and run the system after you leave. Coming soon.
7. EP07 · Make the next deployment easier AI platform engineering Separating customer-specific logic from what the next deployment can reuse. Coming soon.

How is an FDE different from an Applied AI Architect?

One line separates them: who writes code in the customer’s environment. The Applied AI Architect works before the deal, advises across the customer’s stakeholders, and hands over an evaluation and a deployment design. The FDE works after it, sits inside the customer’s team on the customer’s stack, and produces running software. Everything else about the two roles follows from that.

At Anthropic both roles sit inside the same Applied AI organisation, so the handoff happens internally. Some Anthropic postings render the functionally identical role as Applied AI Engineer. That label travels less reliably than the work does: xAI uses Applied AI Engineer for applied product and recommendation engineering with no customer embedding at all. Read the responsibilities, not the title.

What does this mean if you are buying AI?

It means the vendor has accepted that their model does not deploy itself, and has budgeted an engineer to sit in your building until it does. That is a stronger commitment signal than any implementation SLA, and it is worth asking for explicitly during procurement. It also changes who you need on your side: an embedded vendor engineer is only as useful as the internal owner they work with, which is normally an AI Engineer or the Chief AI Officer’s delivery team.

Sources: Rung, Forward Deployed Engineer jobs census (November 2026) · Anthropic, Forward Deployed Engineer, Applied AI · OpenAI, Forward Deployed Engineer (FDE) · Meta, Forward Deployed Engineer, Enterprise AI (August 2026).