Anthropic trains AI engineers like medical residents

Anthropic is putting $100 million behind a plan to train 10,000 engineers at its enterprise customers and consulting partners by the end of 2027, through a program called Claude Frontier Academy that Anthropic announced on October 2. If your company wants a seat, participation is by nomination: the organization asks its Anthropic account team or Partner Account Manager whether it is eligible, then puts forward an engineer who arrives with a named Claude project to lead once back at work.
The candidate Anthropic describes is a hands-on software engineer with strong fundamentals and a record of building with LLMs, who has helped others adopt AI and tends to take on the business’s most important problems. Prior experience building AI agents isn’t required. Organizations are asked to send their strongest engineers. Pricing for nominating organizations has not been announced.
Two exams, two badges
The first program, the Frontier Deployed Engineer Residency, borrows from medical training: residents learn from practitioners, work on realistic cases and are assessed before they practice alone. It runs in two stages, and each stage ends in an assessment.
Stage one is a multi-day in-person program with Anthropic engineers and licensed instructors. It is built around a simulated enterprise deployment that runs from choosing the right use case through security review to handover, and it closes with a graded practical on a scenario the engineer hasn’t seen. Those who pass get the Claude Resident Engineer badge.
Stage two is a 12-week residency at the engineer’s own organization, leading a real Claude use case with support from Anthropic engineers and the rest of the cohort. A second assessment at the end decides who earns the Claude Frontier Deployed Engineer badge. Anthropic expects the first of those in early 2027.
The standard being taught is Anthropic’s own. The Academy trains to “the same standard of skills as Anthropic’s own engineers”, and both badges carry Claude’s name, so what a graduate holds is a credential in one vendor’s stack. Steve Corfield, Anthropic’s Global Head of Business Development and Partnerships, says the goal is for graduates “to set the standard for how AI gets built inside a business.”
Who is in the first cohorts
Cohorts are already running in San Francisco, New York and London, with engineers from Accenture, Bain, Capgemini, Commonwealth Bank of Australia, Deloitte, McKinsey, Morgan Stanley, Novo Nordisk and others. Five of the eight named firms are consultancies that build AI systems for clients. The other three are a bank in Australia, a US investment bank and a drugmaker.
The job title comes from Palantir, which pioneered the forward deployed engineer model in the 2000s. Its engineers were embedded with customers and built whatever was needed, because customer data was scattered across legacy systems and departments. Anthropic’s version, renamed “Frontier Deployed”, trains the customers’ and partners’ own engineers. Accenture’s statement still uses the original term for its staff, describing forward deployed engineers who work “inside client workflows as part of multidisciplinary teams”.
Novo Nordisk’s statement points to Claude Science in its R&D workflows. Novo brought that in on September 16, alongside OpenAI’s GPT-Rosalind, which it had already been using under a deal signed in April. So its engineers will train on Claude while the company runs both vendors’ science models.
Which $100 million
The figure matches another one. On March 12, when it launched the Claude Partner Network, Anthropic committed an initial $100 million to it for 2026. A large share of that money goes straight to partners for training and sales enablement, market development and co-marketing. The Academy’s $100 million is a separate commitment tied to one goal, 10,000 engineers, with a deadline at the end of 2027.
The Academy sits above what the Partner Network already offers. Professionals at 46,000 firms have earned more than 175,000 Claude certifications, starting with Claude Certified Architect, Foundations, an exam for solution architects that arrived in March. Nearly 4,000 people have finished Basecamp, an immersive course that uses the curriculum Anthropic’s Applied AI teams get when they join. The residency’s target is more than twice Basecamp’s total so far, and every one of those engineers has to pass two graded assessments.
Capgemini’s statement in the announcement was blunt about the difference: “What I’m hearing from clients isn’t about certification counts or training volume; it’s about knowing firsthand how to solve real-world problems that create value to our clients.”
Up next
OpenAI’s GPT-6 guide says drop blanket “always ask” rules
Teams running GPT-6 agents on jobs that last hours or days can send a correction partway through. OpenAI’s guide says the correction waits in a queue and doesn’t stop tools that are already running.
Related stories
- At least $1B for outside evaluators inside Anthropic
- Anthropic’s $150M buys Claude time for federal scientists
- Ozempic maker adds Claude next to OpenAI’s Rosalind
- Claude Marketplace adds CrowdStrike, Cursor and Vercel
- Indirect channels now carry 40% of Anthropic’s ARR
- Anthropic’s IPO math leans on a $190B to $200B revenue forecast for 2028
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

