Anthropic announced on 2 October a $100 million programme to train 10,000 “Frontier Deployed Engineers” by the end of 2027 — and the first cohorts read like a list of companies with big workforces in this corner of the world: Accenture, Bain, Capgemini, Deloitte, McKinsey, Morgan Stanley, Novo Nordisk and the Commonwealth Bank of Australia are all in. The company’s diagnosis is blunt: what is holding enterprise AI back is not model capability. It is the shortage of people who can install it and make it work.
🔍 THE BOTTOM LINE: The largest training bet by a frontier AI lab so far treats AI deployment skill — not coding and not model-building — as the scarce resource of the AI economy, and it is being built with the same residency model used to train doctors. For workers across Australia and New Zealand, the programme’s shape is as interesting as its size: the jobs it trains for sit at banks, consultancies and industrials, not just Silicon Valley.
The residency model, applied to AI
The Frontier Deployed Engineer Residency works like a compressed medical training pathway. Engineers nominated by their employers spend a multi-day in-person intensive building a Claude system for a simulated enterprise, finishing with a graded practical. Pass, and they return to their own organisation for a 12-week residency leading a real AI deployment, supported by Anthropic engineers. A final assessment earns the Frontier Deployed Engineer credential, with the first badges expected in early 2027.
Two details stand out. Experience building AI agents is explicitly not required — the programme targets hands-on engineers with strong fundamentals. And participation is by nomination only, routed through Anthropic’s account teams, which tells you who the programme is really for: the enterprises already paying for Claude at scale.
A bank with 20 seats and 5 million customers across the Tasman
Commonwealth Bank of Australia said on 3 October that an initial group of 20 engineers will take part — notable for New Zealand readers because CBA owns ASB, one of this country’s major banks, and runs AI programmes across both markets. A bank training twenty engineers to deploy agentic AI inside its own operations is a workforce signal that travels across the Tasman faster than any press release.
The consulting firms in the first cohorts — Accenture, Deloitte, Bain, Capgemini, McKinsey — all operate large New Zealand practices. Training their deployment engineers offshore, then stationing them in client work here, is a plausible path by which this programme reaches NZ employers without any local announcement at all.
The numbers behind the “people bottleneck” claim
Anthropic’s bet lines up with independent data published the same week. Talent-intelligence firm Draup’s Macro Labour Market Trends analysis, released 30 September, found AI-specific “builder” roles now claim 27 percent of Fortune 500 tech postings, and that forward-deployed style roles have grown into their own family — while Anthropic’s own announcement claims “a small group of deeply skilled people drives an outside share of what AI delivers” inside its client companies.
The framing also fits what this site has tracked all year. OpenAI’s deployment subsidiary is hiring hundreds of forward deployed engineers, and the FDE role was already the fastest-growing job title in AI months before Anthropic moved. What is new here is not the role — it is the scale of the training infrastructure being stacked underneath it.
What it means for workers who are not in the first cohort
The programme is closed for now, but its curriculum is public by design: use-case selection, security review, simulated deployment, then a graded handover into production. Those are learnable skills, and several appear in the free and low-cost pathways we track in the Career Compass resource list. Meanwhile, NZ job ads mentioning AI skills have kept climbing faster than overall hiring, and most Kiwi workers in recent survey data say AI is helping them build more valuable skills — the same self-taught instinct this programme is formalising.
The caveat worth holding: this is an AI lab training workers on its own product, assessed by its own instructors, with a badge that is only meaningful while Claude remains central to enterprise AI. It is a workforce programme, but it is also a go-to-market strategy. Both can be true, and readers weighing a career bet should weigh both.
🔍 THE BOTTOM LINE: A $100 million training programme will not, on its own, fill the AI skills gap — but it does tell you where the industry thinks the gap actually is. Not in research labs, and not in model weights: in the interface between AI systems and the businesses that need them working. New Zealand firms, long priced out of frontier-lab attention, should watch whether the big consultancies bring what their engineers learn back across the Tasman.
❓ FAQ
What is a Frontier Deployed Engineer? An engineer who deploys AI systems inside a real organisation — choosing use cases, managing security review, and seeing a working system through to handover. Anthropic estimates this is the talent most in shortage for enterprise AI.
How long does the Anthropic Frontier Academy programme take? About three months: a multi-day intensive with a graded practical, then a 12-week residency leading a live Claude deployment at the participant’s own employer.
Can individuals apply, or only companies? Only organisations can nominate participants. Anthropic directs interested companies to its account or partner teams; there is no open enrolment.
Does the programme help workers outside the first cohorts? Indirectly. The curriculum and the credential’s existence signal which skills are scarce, and competitors are likely to respond with training of their own — as OpenAI already has with its deployment-company hiring push.
📰 Sources
- Anthropic — Claude Frontier Academy announcement
- NeoTeo — Commonwealth Bank participation reporting
- ODSC — The Week in AI Careers, Sept 28–Oct 4
- Draup — Macro Labor Market Trends 2026
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
