When people picture AI hiring, they tend to picture research scientists. The actual growth is happening somewhere else: a job spent mostly on planes, in client offices, making other companies’ half-finished AI projects work.
OpenAI’s careers site lists 833 open roles, and a large and growing share of them carry the same title: Forward Deployed Engineer. The role is the core of the OpenAI Deployment Company, a majority-owned subsidiary launched on 11 May 2026 with more than $4 billion from a 19-firm consortium led by TPG, seeded by the acquisition of Edinburgh firm Tomoro and its roughly 150 experienced deployment engineers.
The title is barely five years old as a mass-market job category. The hiring data behind it is the clearest “AI is creating jobs” story of 2026 — and one of the strangest new career paths the industry has produced.
What a forward deployed engineer actually does
The role, invented by Palantir and named Forward Deployed Software Engineer, sends engineers to sit inside a customer’s organisation — not to sell software, but to build it, against the customer’s real data, legacy systems and office politics, staying until it works.
OpenAI’s description of the job is blunt about the scope: FDEs “work closely with business leaders, operators, and frontline teams to identify where AI can make the biggest impact, redesign organizational infrastructure and critical workflows around it, and turn those gains into durable systems.” That is part engineering, part change management, part organisational surgery.
The economics explains why the labs are paying up. According to a Plank Research census from September 2026, national postings for the role grew more than 1,000 percent from January through August 2026 versus the same eight months of 2025, while broader tech postings grew 13 percent. Plank counted 1,333 live postings across 565 companies — still small in absolute terms, which is precisely why the pay is aggressive: the supply of people who can do it is structurally thin.
The money, and what it says about demand
A compensation report by Perspective AI, synthesising 1,200 data points from Levels.fyi, public postings and pay-transparency bands, puts median mid-level total compensation for the role at $385K, staff level at $610K, with principal FDEs at frontier labs clearing $1.2M — against a Palantir FDSE median of about $215K. Bay Area-specific reporting cited by Bloomberg put median advertised salary at $202,300, with OpenAI postings ranging roughly $145K-$325K plus equity and Anthropic advertising $280K-$320K.
Two structural notes deserve the same airtime as the headline figures. First, the equity share at frontier labs now runs 55-70 percent of total compensation, up from 35-45 percent in 2024 — meaning the eye-catching numbers are correlated bets on AI valuations holding, not cash. Second, Fortune 500 enterprise AI teams — where most AI-engineer titles actually live by headcount — pay $190K-$420K with modest equity, and big banks are now base-competitive with the labs at staff level. The premium is real but concentrated.
And the title inflation cuts both ways. Job boards carry “FDE-flavoured” roles — solutions architects, applied AI engineers, deployment leads — that make any census of the category fragile. Bloomberg-adjacent reporting has been careful to note the absolute numbers remain modest; a hot title with 1,333 national postings is a boom at the scale of a mid-sized company’s workforce, not a mass market. Not yet, at least.
The demand concentrates at the labs, and it is visible on their own boards. OpenAI’s careers site currently lists 833 open roles, and the GAGE hiring index — a weekly count from 19 large AI employers’ public boards — recorded 60 open governance-titled roles at OpenAI and 47 at Anthropic in its 28 September snapshot, a reminder that even the “control the technology” jobs sit disproportionately at the labs. For job-seekers the practical read is that the labs are hiring deployment and governance people faster than almost anything else, while the volume tier of AI-era hiring grows at large enterprises paying less but hiring more.
The hiring boom behind the boom
The Deployment Company matters for jobs arithmetic beyond its own headcount. OpenAI committed to using its roughly $4 billion partly to “acquire firms” — an explicit consolidation plan for an industry fragment of consultancies, data shops and deployment boutiques. Its investor partners — TPG, Bain Capital, Brookfield, Goldman Sachs, SoftBank Corp, plus Bain & Company, Capgemini and McKinsey — collectively sponsor more than 2,000 businesses, creating a distribution channel that converts consulting relationships into engineering headcount.
The pattern is visible in the wider data. A study released 21 September by Stanford’s Bharat Chandar and King’s College London’s Bouke Klein Teeselink, covering 1.25 billion job postings and 154 million employment records across 41 countries, found AI-adopting companies added more jobs overall than non-adopters, growing senior headcount 6.7 percent over five years while junior hiring fell 3 percent. Employment in computer and mathematical occupations rose 0.8 percentage points at AI adopters. The new jobs exist — they skew experienced.
That is the pattern in one job title. The forward deployed engineer is, almost by definition, a senior role: the hiring bar, per the postings, demands production engineering plus customer-facing judgement that normally takes years to build. LinkedIn’s Jobs on the Rise data has shown AI engineering roles dominating fast-growth lists while entry-level hiring thins — and the FDE wave is the senior end of that same reshuffle.
The honest caveats
Three things complicate the upbeat read, and they belong in any jobs story about this market.
It is early and it is small. A role with roughly 1,300 live national postings is a rounding error against US tech employment. The 1,000 percent growth rate is off a tiny base — genuinely fast, genuinely new, and genuinely narrow.
The same companies are cutting elsewhere. Bay Area tech layoff notices covered more than 14,500 workers in the year to June 2026, nearly double the prior year, led by Meta’s 3,715 planned cuts, even as San Francisco’s total AI-sector hiring grew. San Francisco’s chief economist Ted Egan told Bloomberg the pattern is “a lot of people getting laid off and a smaller number of people… getting hired for AI-specific skills.” Both things are happening, and pretending otherwise would be the kind of cheerleading this field does not need.
The services business has a cyclical history. FDE work is, structurally, high-end consulting wearing an engineering badge. Consultancy is cyclical, and a slowdown in enterprise AI spending would land on exactly these teams. The Perspective report itself recommends candidates model realisable compensation at a 40 percent valuation drawdown — unusual candour for a hiring-boom story.
What it means for the rest of the market
The transferable part of the story is not the $385K median — it is the description of what AI-era employers pay for: people who can arrive at an organisation that bought an AI tool, work out why it is not working, redesign the workflow around it, and stay until the thing is trusted. That skill mix — technical enough to build, senior enough to persuade — is exactly what SEEK’s New Zealand data has shown appearing in job ads here too, where AI ethics and governance is the fastest-growing named AI skill. New Zealand will not host a frontier lab’s deployment army, but the same reshaping is visible locally: fewer ads asking for pure model-building, more asking for people who can make AI work safely inside a real organisation.
The strongest, most defensible jobs claim of 2026 sits in the middle ground the loud narratives keep missing: AI is simultaneously shrinking some rungs of the career ladder and creating, from nothing, a job category that pays like elite special forces. Both statements are in the data. Workers deserve to see both.
FAQ
What is a forward deployed engineer? An engineer embedded inside a customer’s organisation to build and ship AI systems against the customer’s real data and workflows — a role Palantir invented and OpenAI, Anthropic, Google and Microsoft now hire for aggressively.
How fast are forward deployed engineer postings growing? Plank’s September 2026 census counted more than 1,000 percent national posting growth (January-August 2026 vs the same period in 2025), while broader tech postings grew 13 percent — off a small base of roughly 1,333 live postings.
How much do forward deployed engineers earn? Median advertised salary is around $202,300 in Bay Area data; compensation reports put mid-level total comp at about $385K and senior frontier-lab FDEs above $780K, largely through equity.
Can people outside the US get these roles? OpenAI, Anthropic and Google post FDE roles across the UK, Singapore, Japan and Australia. Reported international bands typically run 50-70 percent of US compensation.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
