On October 2, 2026, Anthropic announced Claude Frontier Academy, a $100 million commitment to train 10,000 "Frontier Deployed Engineers" by the end of 2027. The term borrows from the forward-deployed engineer model, where a vendor's engineers sit inside a customer to get a system working. The difference here is that the engineers being trained work for the customers and partners themselves.
How the program works, per Anthropic's announcement:
- Phase one: a multi-day, in-person program led by Anthropic engineers and licensed instructors. Participants work through a simulated enterprise deployment, from choosing a use case to passing a security review, then take a graded practical assessment on a new scenario. Passing earns a "Claude Resident Engineer" badge.
- Phase two: a 12-week residency in which the engineer leads a real Claude project at their own organization, with support from Anthropic engineers and their cohort. A final assessment earns the "Claude Frontier Deployed Engineer" badge. Anthropic expects the first of these in early 2027.
- Who: nomination only, through an organization's Anthropic account team or partner manager. First cohorts run in San Francisco, New York and London, with engineers from Accenture, Bain, Capgemini, Commonwealth Bank of Australia, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk. The announcement does not say what participants or their employers pay.
How teams are using it. The mix of the first cohort is telling. Five of the eight are consultancies and integrators, which will use certified engineers to staff client deployments. The other three are a bank, an investment bank and a pharmaceutical company, which will use them to run projects in-house. In both cases the residency puts the training on a live system that touches the organization's own data, permissions and security review, not a sandbox. Anthropic's stated reason is that implementation talent is "the hardest talent to find," and that small skilled teams deliver outsized value.
Enterprise pattern (analysis): the design reflects where enterprise AI projects actually stall. The model is rarely the problem; scoping the use case, getting through security review, building evaluations and handing the system to an operations team usually are. A program that grades engineers on a simulated security review is training for that constraint directly. The open question is lock-in: the skills are mostly transferable, but the badge, the playbook and the support channel are all Claude-specific, and a company that builds its deployment team around one vendor's certification will find switching more expensive.
What to watch: whether completion rates and project outcomes are published, whether rival labs answer with their own certification tracks, and how many of the 10,000 places go to consultancies rather than end customers.
Anthropic's $100M Claude Frontier Academy aims to certify 10,000 engineers by end-2027 through a deployment simulation plus a 12-week residency on a live in-house project, targeting the real enterprise bottleneck -- scoping, security review and operations -- while deepening customers' reliance on Claude-specific skills.