Summer 2027 recruiting
Anthropic internships and the Fellows Program
No internships: Anthropic's careers page says so outright. The nearest thing to a student program is the Anthropic Fellows Program, a four-month, full-time funded research fellowship with Anthropic mentors. Everything else is regular hiring, and Anthropic says it cares about what you can do, not where you learned it.
AnthropicEarly-career programs
- Anthropic Fellows Program, AI Safety & Security
- Anthropic Fellows Program, ML Systems & Reinforcement Learning
- Anthropic Fellows Program, The Anthropic Institute (Economics & Policy)
Internships and programs at Anthropic
The early-career programs it runs for students and recent graduates, by the names it uses for them.
- Anthropic Fellows Program, AI Safety & Security
- Four months of full-time empirical research with Anthropic mentors, a weekly stipend of 3,850 USD (or 2,310 GBP or 4,300 CAD), and about $15,000 a month in compute. Safety projects cover scalable oversight, adversarial robustness and AI control, model organisms, interpretability, and AI welfare; security projects cover vulnerability research, red teaming, and control evaluations. Fellows work from shared spaces in Berkeley or London, or remotely in the US, UK, or Canada.
- Anthropic Fellows Program, ML Systems & Reinforcement Learning
- A parallel track for people whose strengths are on the systems and reinforcement learning side. Same structure as the safety track: a funded, mentored four-month research project with the same location options.
- Anthropic Fellows Program, The Anthropic Institute (Economics & Policy)
- A track for research on the economic and policy side of AI. It fits applicants from economics, policy, or quantitative social science who want hands-on research with Anthropic mentors.
When Anthropic posts internships
No internship means no fall intern rush to plan around. Anthropic runs several Fellows cohorts a year and reviews applications as they arrive; the current listings advertise a January cohort, and an earlier round set a late-July deadline for a November start.
Regular early-career roles post whenever a team has an opening, with no campus calendar, so there's no single month when student-friendly jobs appear.
These are patterns from recent cycles, not published dates, and they shift. The reliable signal is the posting itself: Trakker is an internship tracker that reads Anthropic’s careers site directly and can email you within minutes of a new role appearing.
Who can apply
Fellows need work authorization in the US, UK, or Canada and must live in that country during the program; Anthropic doesn't sponsor visas for it. A PhD, prior ML experience, and publications are not required, but fluent Python and a strong technical background (computer science, math, physics, cybersecurity, or another quantitative field) are expected.
It's full-time, 40 hours a week for four months, so it fits a gap semester, a leave, or the months after graduation far better than a normal school term. Fellows are funded participants, not employees. Anthropic says 25 to 50 percent of fellows in previous cohorts received a full-time offer, with no guarantee.
The interview process
The Fellows process runs through an application and reference check, technical assessments and interviews, and a research discussion, then project selection and mentor matching. Constellation, Anthropic's recruiting partner, manages applications and interviews.
For regular technical roles, interviews happen over Google Meet using live coding tools like Colab and CodeSignal, and you're allowed to look things up. Non-technical interviews are conversational: how you approach problems and why you care about the mission.
Tips for applying to Anthropic
- The Fellows application rewards evidence that you can turn a vague research question into running experiments. Link a repo or write-up where you did exactly that.
- Check the work-authorization rule first. Without US, UK, or Canadian authorization and residence during the program, you can't do the fellowship.
- About half of Anthropic's technical staff had no prior ML experience, by the company's own count. A strong engineering record in another area is a real path to a full-time role later.
Anthropic internship questions
- When does Anthropic post internships?
- It doesn't. Anthropic's careers page states that it does not currently offer internships. The closest option for students and recent graduates is the Anthropic Fellows Program, which runs several cohorts a year.
- What is the Anthropic Fellows Program?
- A four-month, full-time funded research fellowship: fellows work on an empirical project with Anthropic mentors, a weekly stipend, and a compute budget. Current tracks cover AI safety and security, ML systems and reinforcement learning, and economics and policy.
- Do you need a PhD for the Anthropic Fellows Program?
- No. Anthropic says a PhD, prior ML experience, and publications are not required. You do need fluent Python, a strong quantitative background, and work authorization in the US, UK, or Canada.
- What is Anthropic's interview process like?
- Technical roles use live coding in tools like Colab or CodeSignal over Google Meet, and candidates can look things up. The Fellows Program adds a research discussion and a mentor-matching step after the technical interviews.
Where else to look
Employers students often weigh against Anthropic, and the role tracks its internships fall under.
Sources
What this page was checked against. Programs and timelines change from cycle to cycle, so Anthropic’s own careers site has the final word.
Hear about Anthropic the minute it posts.
Follow Anthropic and hear about every internship and new grad role it puts up: within minutes on a paid plan, or in the 7pm ET digest on the free one.
You won’t need a card. Used by hundreds of students.