[Anthropic](https://www.anthropic.com)

April 26, 2026 (application deadline for July 2026 cohort)

Weekly stipend of $3,850 USD (or equivalent in GBP/CAD) for 4 months; ~$15,000/month for compute and research expenses

Supports engineers and researchers to conduct mentored, high-impact empirical AI safety research with public outputs.

Applicants must have work authorization in the US, UK, or Canada; no visa sponsorship available.

--- ## Description The Anthropic Fellows Program is a four-month, full-time, funded opportunity for engineers and researchers to work on high-priority AI safety research questions. Fellows are paired with Anthropic mentors and collaborate on projects that align with Anthropic’s research priorities, such as scalable oversight, adversarial robustness, AI control, mechanistic interpretability, AI security, and model welfare. The program emphasizes producing public research outputs (e.g., papers, open-source tools) and offers hands-on experience, mentorship, and networking within the AI safety community. Past fellows have published on topics including agentic misalignment, subliminal learning, rapid response to AI jailbreaks, and open-source interpretability tools. The program is designed to foster new talent in AI safety, regardless of prior experience, and previous cohorts have seen a significant proportion of fellows transition to full-time roles in the field. --- ## Due Dates - April 26, 2026: Application deadline for the cohort starting July 20, 2026 Applications are also accepted on a rolling basis for future cohorts, but to be considered for the July 2026 start, apply by April 26, 2026. --- ## Funding Amount - Weekly stipend: $3,850 USD / £2,310 GBP / $4,300 CAD (varies by country) - Duration: 4 months (full-time) - Additional funding: Approx. $15,000 USD per month for compute and other research expenses - Benefits may vary by location --- ## Eligibility - Open to technical researchers and engineers with strong fundamentals in Python and quantitative fields (computer science, mathematics, physics, etc.) - Must have full-time work authorization and be physically located in the US, UK, or Canada during the program - No visa sponsorship is available; candidates must secure their own work authorization if needed - Open to candidates from a variety of backgrounds; a PhD or prior ML experience is not required - Strong motivation for AI safety research, ability to learn quickly, and drive projects forward are key selection criteria - Fluency in Python programming is required --- ## Application Process - Submit an application via the [official Constellation application form](https://bit.ly/afpsafety) (Anthropic’s recruiting partner) - Application includes initial form, reference check, technical assessments/interviews, and a research discussion - Applicants can indicate workstream preferences (AI Safety, AI Security, ML Systems & Performance, Reinforcement Learning, Economics & Societal Impacts) - Applications are reviewed on a rolling basis for future cohorts, but April 26, 2026 is the deadline for the July 2026 cohort - Strong candidates are encouraged to apply even if they do not meet every listed qualification --- ## Additional Information - Fellows may work remotely (within the US, UK, or Canada) or in shared workspaces in Berkeley, CA or London, UK - No guarantee of a full-time position after the fellowship, but strong performance may lead to offers (historically 25-50% of fellows received offers) - The program is highly collaborative and values diversity of perspective - For more details on past projects and research directions, see the [Alignment Science Blog](https://alignment.anthropic.com/) --- ## External Links - [Program Details & Application Portal](https://job-boards.greenhouse.io/anthropic/jobs/5023394008) - [Anthropic Fellows Program Blog Post](https://alignment.anthropic.com/2025/anthropic-fellows-program-2026/) - [Alignment Science Blog](https://alignment.anthropic.com/) - [Anthropic Research](https://www.anthropic.com/research) - [Apply via Constellation](https://bit.ly/afpsafety) - [Anthropic Careers](https://www.anthropic.com/careers) --- ## Contact Information | Contact Type | Details | |------------------|-------------------------------------------------------| | General Contact | [email protected] | | Online Contact | [Anthropic Contact Form](https://www.anthropic.com/contact) | | Recruiting Partner | [Constellation](https://www.constellation.org/) | ---

Anthropic Fellows Program for AI Safety Research

Sign in to see full grant listing