What a Forward Deployed Engineer Does (and What It Pays)

Welcome back to what each role does, our series on AutomationRoles.ai that takes one fast-growing job in automation and AI and explains what the work actually involves, who's hiring, what it pays and how to break in.

Last time we looked at data centre robotics engineers. This time it's a role I think is one of the best-kept secrets in tech careers right now: the forward deployed engineer, or FDE.

If you've read our October hiring roundup, you'll remember that forward-deployed and applied engineering was one of the biggest technical job families at AI companies. Now we've got much sharper data on it, and it's worth a proper look.

What a forward deployed engineer actually does

Here's the simplest way I can put it. A model on its own doesn't change a business. Someone has to sit with the customer, work out what problem is really worth solving, wire the AI into their messy data and systems, prove it works, and then keep it working. That someone is the forward deployed engineer.

OpenAI's own FDE posting describes the team as partnering "with customers to turn research breakthroughs into production systems", and says FDEs "embed with customers where model performance matters, delivery is urgent, and ambiguity is the default." I love that line, because it's honest. This is not a tidy job.

Sayari, which builds risk and supply chain intelligence for government agencies and large enterprises, puts it even more bluntly in its FDE II posting: "The same engineer scopes the problem and ships it." It describes the person as "equal parts engineer, strategist, and operator", and says "You prefer a working prototype to a slide about one."

That's the job in one sentence. You are a real software engineer, but you live at the front line with the customer instead of deep inside a product team.

A week in the life

Pulling together the October 2026 postings I read, an FDE's work typically looks like this:

• Discovery. Sitting with customer engineers, analysts and executives to understand their workflows, constraints and what success actually means.
• Proving it. Running demos, proofs of concept and trials on the customer's own data. Sayari lists "win the technical evaluation" as the first thing you'll do.
• Building production software. Not slides, not prototypes that die. The AWS FDE II posting for US Government work talks about retrieval pipelines, workflow automation, API orchestration layers and data integration for "AI and agentic workflows".
• Evaluation and guardrails. Designing experiments, graders and benchmarks to show whether an approach really works. AWS lists monitoring, evaluations, guardrails and CI/CD; OpenAI asks for "reproducible experiments, evaluations, graders and benchmarks".
• Owning production. Troubleshooting incidents across models, distributed systems and application services, and staying accountable until it's handed over.
• Feeding the product. Deciding what's bespoke, what's reusable and what belongs in the core platform, then turning field lessons into tools and playbooks for the next engagement.

Buildkite, the CI/CD company with its engineering team in Australia, sums up the commercial heart of it in its FDE posting: "Customers pay for an outcome: you work out what they actually need, agree on a clear deliverable, build it as production code, and land it in the platform." It adds that there's no commission: "you're measured on outcomes delivered, not on what the customer spends."

That's what makes this role so interesting to me. It's engineering with real customer access, real stakes and very fast feedback.

How big is the FDE market?

Bigger than most people think. FDE Dispatch took a snapshot of public job boards (Greenhouse, Lever, Ashby and Workday) on 24 October 2026 and found 725 live FDE postings across 121 employers. Of those, 568 had engineer titles, 86 were deployment strategist roles and 71 were FDE management or program roles.

The biggest posters in that snapshot were:

• Databricks: 122 postings
• Palantir: 105
• Salesforce: 50
• Accenture: 40
• OpenAI: 32
• LangChain: 17
• Cognition: 15
• ElevenLabs: 12
• Cloudflare, Cohere and Scale AI: 11 each

The top five employers account for 48.1% of postings, so it's concentrated, but there's a long tail too: forty employers had exactly one FDE opening. That long tail is where I'd be looking if you want to be one of the first FDEs at a growing company and help shape how the function works.

One honest caveat, which FDE Dispatch states itself: these are open postings, not headcount, and employers that only post on LinkedIn or other systems aren't counted. So treat it as a solid lower-bound signal rather than a census.

What it pays

This is where people sit up. Across 255 US engineer-title FDE postings with posted pay, FDE Dispatch's salary analysis found a median posted range of $180,000 to $250,208, with a median midpoint of $216,104. Counting each employer once, the median midpoint was $208,000.

Some real examples from postings I checked or that FDE Dispatch verified by hand on 22 October 2026:

• OpenAI, Forward Deployed Engineer (FDE), AI Infra, San Francisco: $185K–$300K plus equity, with travel up to 50%.
• Buildkite, Forward Deployed Engineer, US West Coast: $223,000–$247,000 base plus equity, around 25% travel.
• Sayari, Forward Deployed Engineer II, remote US: $160,000–$180,000, with 25–30% travel and in-office time.
• Anthropic, Staff Forward Deployed Engineer (New York, San Francisco, Seattle): $280,000–$320,000, as listed by FDE Dispatch.
• Palantir, Forward Deployed Software Engineer, New York: $135,000–$200,000, which may also include RSUs and a sign-on bonus.
• Scale AI, Frontier Agents Engineer (Forward Deployed Engineering): base $180,000–$225,000 in SF, NY and Seattle.

Outside the US, posted pay is rare (FDE Dispatch found only 6.9% of non-US postings listed it), but the one London example in its verified table, an Anthropic FDE role, listed £225,000–£255,000.

Remember these are posted ranges, not offers. Location, seniority, clearance requirements and equity all move the real number.

Where the jobs are

By city, FDE Dispatch's location breakdown puts New York (152 postings from 56 employers) and the San Francisco Bay Area (148 from 64) out in front, followed by London (77) and the Washington, DC area (75). Outside the US, Singapore, Munich, Sydney and Toronto each had 15 to 22 postings.

Sydney showing up with 15 postings from 8 employers, including Accenture, Salesforce, Palantir, Databricks and Cognition, makes me smile. Australia and New Zealand are often treated as an afterthought in AI hiring, and this is a role where being in the same time zone as the customer genuinely matters.

On remote work: 27.4% of postings listed a remote option, but Databricks accounted for 95 of those. Strip Databricks out and it's 17.2%. And whatever the listing says, expect to travel. 59.3% of postings mention travel, with 20%, 25% and 50% the most common stated figures.

Who succeeds in this role

Reading across the postings, the profile is remarkably consistent:

• Production engineering chops. Sayari wants production Python, Pandas, SQL and REST APIs, plus cloud and containers (Docker, Kubernetes). OpenAI wants people who "write and review production-grade software, and debug unfamiliar systems".
• Hands-on LLM and agent experience. OpenAI asks for experience building or deploying systems powered by LLMs; Sayari lists orchestration, tool use, retrieval and evaluation as desired.
• Customer-facing experience. Both Sayari and OpenAI ask for around 5+ years that include client-facing technical work. Sales engineers, solutions architects and technical consultants: this is your door.
• Comfort with ambiguity. Every posting says it in some form. You'll parachute into an unfamiliar company, learn their stack fast and make decisions with incomplete information.
• AI-assisted coding as a habit. Sayari explicitly lists "AI-assisted coding tools (Cursor, Claude, or similar) as a daily part of how you work." Expect that to become standard.

Notice what isn't required: a PhD. This is one of the most accessible high-paying AI roles for strong, pragmatic engineers.

How to break in: my practical advice

1. Build one real deployment, end to end. Pick a messy, real-world problem (a small business's invoices, a sports club's bookings, your own team's support inbox), connect an AI agent to real data, add an evaluation harness, and measure the result. Use whatever model or assistant you're comfortable with, whether that's Grok Bot, Claude, OpenAI's models or an open-source stack. The point is the outcome, not the tool.
2. Write it up like a consultant. Problem, success criteria, architecture, what failed, what you'd do next, time or money saved. That write-up is your FDE interview in miniature.
3. Learn evals properly. Graders, baselines, regression tests for prompts and agents. Almost every FDE posting mentions evaluation, and it's the skill that separates demos from production.
4. Practise the customer conversation. If you've never run a discovery session, volunteer to. Sit in on sales calls. Learn to explain trade-offs to an executive in two minutes.
5. Look at the long tail. Databricks and Palantir hire in volume, but forty employers had a single FDE opening. Being the first or second FDE at a growing AI company is a fantastic way to build a career fast.
6. Get comfortable with travel. If 25–50% travel suits your life right now, you have an edge. If it doesn't, filter hard for the roles that say so upfront.

My take

I'm genuinely excited about this role. For years the gap in tech hasn't been the technology, it's been getting the technology to work inside a real business. The forward deployed engineer is the person who closes that gap, and the market is now paying properly for it.

It's also, I think, one of the most future-proof jobs in automation. As models get more capable, the value shifts even further towards people who understand the customer, the data and the workflow, and who can ship something that actually works on Monday morning. That's not a job AI takes away. It's a job AI creates.

If you're a strong engineer who likes people, likes variety and likes seeing your work land, this could be your next move. Start by browsing the latest automation and AI jobs on AutomationRoles.ai and search for "forward deployed" and "deployment".

Sources: FDE Dispatch job titles, salary and location analyses (snapshot 24 October 2026); OpenAI careers, Forward Deployed Engineer (FDE), AI Infra; Sayari Forward Deployed Engineer II on Greenhouse (published 26 October 2026); Buildkite Forward Deployed Engineer on Greenhouse; AWS Forward Deployed Engineer, US Government on Amazon.jobs (October 2026).