Anthropic trains AI engineers like medical residents

Anthropic wants to train AI engineers the way hospitals train doctors: learn from practitioners, work on real cases and pass an assessment before working alone. The model sits behind the Claude Frontier Academy, announced by Anthropic with a $100 million commitment and a target of 10,000 Frontier Deployed Engineers by the end of 2027.

At a glance

• Companies such as Accenture, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk nominate their strongest engineers, who join first cohorts already running in San Francisco, New York and London.
• The path runs from a multi-day in-person program and graded practical to a 12-week residency on a real Claude project, with a badge after each assessment and the first final badges expected in early 2027.
• Anthropic does not say how the $100 million is split, what participating firms pay or how many engineers are expected to pass, so the 10,000 target comes without a pass rate.

If you have not been following Anthropic's partner business, here is the setup. According to Anthropic, it launched the Claude Partner Network in March with a $100 million investment in partner training, technical support and shared marketing. The academy announcement says professionals across 46,000 firms in that network have earned more than 175,000 Claude certifications, and nearly 4,000 people have completed Basecamp, which teaches the curriculum Anthropic's Applied AI teams receive when they onboard.

Two assessments and two badges stand between a nominee and the FDE credential

The academy's first program is the Frontier Deployed Engineer Residency. Anthropic models it on medical training, where doctors learn from practicing physicians, train on real cases and are assessed before they practice alone. FDE residents learn from practitioners, practice on realistic cases and earn the credential only after their work is assessed.

It starts with a multi-day in-person program run by Anthropic engineers and licensed instructors. Participants work through a simulated enterprise deployment, from choosing the right use case through security review to handover, then take a graded practical on a new scenario. Those who pass earn the Claude Resident Engineer badge.

The 12-week residency comes next. Each engineer leads a real Claude use case at their own organization, supported by Anthropic engineers and learning from the cohort while they build. A second assessment at the end decides the Claude Frontier Deployed Engineer badge, and the first ones are expected in early 2027.

Eight organizations are named in the first cohorts, from McKinsey to Commonwealth Bank of Australia

The named first-cohort organizations are Accenture, Bain, Capgemini, Commonwealth Bank of Australia, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk, with others unnamed. Participation is by nomination, and organizations ask their Anthropic account team or Partner Account Manager whether they are eligible. According to CRN, the residency also targets solution providers of various sizes, including boutiques, beyond the global systems integrators.

The bar for nominees is specific: hands-on software engineers with strong fundamentals, a record of building with LLMs, experience helping others adopt AI and a habit of tackling their business's most important problems. Prior experience building AI agents is not required. Each nominee arrives with a named Claude project to lead after returning.

The partner statements add a few details. Commonwealth Bank of Australia says its teams produced up to 3x more code changes in the past year with AI tools. Novo Nordisk says it already uses Claude Science in its R&D workflows. Capgemini argues that clients care about solving real problems more than about certification counts or training volume.

Anthropic's Services Track tiers start at 10 certified people and go up to 1,000

Steve Corfield, Anthropic's Global Head of Business Development and Partnerships, frames the problem as depth of talent. In his telling, a small high-agency team with Claude access and a deep grasp of how the business runs can transform a company, and the hard part is building engineers who can take Claude from an idea to a system in production.

The partner program puts a number on that talent. According to Anthropic, the Services Track tiers partners by certified bench: Select needs at least 10 active certified individuals, Preferred at least 100, and Global Premier at least 1,000 plus at least 100 deployed joint customers across three or more regions. Anthropic also reports that Accenture is training 30,000 professionals, while Cognizant has rolled Claude out to roughly 350,000 associates.

By the same Anthropic account, Deloitte is making Claude available to 470,000 people and KPMG is integrating it across a workforce of more than 276,000. According to CRN, Anthropic expects the academy to expand beyond engineering to executives and knowledge workers. The job title is already contested: according to Palantir, its own «AI FDE» is a software agent that turns natural language requests into Foundry operations.

The residency grades a whole deployment, from use case to handover

The design choice worth noticing is what gets graded. A typical certification checks whether you know the tools. The residency practical drops you into a new simulated enterprise scenario and grades the whole chain: picking the use case, getting through security review and handing the system over to the people who will run it.

Think of the gap between a written driving theory test and a road test on streets you have never driven. Accenture describes the residency as judgment tested against a simulated enterprise deployment, not technical skills in isolation. According to Anthropic, the reasoning is that a successful pilot is not a system a business can run on, because the real work sits in integration, evaluation and how people's work changes.

The announcement leaves the economics blank. It does not say how the $100 million is spent, whether nominating firms pay or what share of engineers pass either assessment. In our view, 10,000 looks like a tight target for this format, since each final badge requires a 12-week residency supported by Anthropic engineers and the first badges arrive only in early 2027, about a year before the deadline.

When the first FDE badges land. The first Claude Frontier Deployed Engineer badges are expected in early 2027. That will be the first chance to see how many residents get through both assessments. The program is set to grow toward 10,000 engineers by the end of 2027, and Anthropic calls that the first step of many. According to CRN, programs for executives and knowledge workers will follow, but no dates for them have been given.