Fellowship
Open
Anthropic Residency 2026: paid AI safety research placements
Anthropic Residency pays technical leaders $350,000 to $500,000 a year equivalent for AI safety research placements. Apply by 22 November 2026.
| Detail | Value |
|---|---|
| Placement length | 6 weeks, 3 months or 6 months; other lengths between 3 and 6 months by exception |
| Hours | Full time, flexible start dates |
| Remote work | Possible, mainly for shorter placements; San Francisco office preferred |
| Support | One to one Anthropic mentor; self study time and talks; access to frontier models and a compute budget |
| Research areas | Scalable oversight; adversarial robustness and AI control; model organisms; interpretability; safeguards; AI welfare; AI security and red teaming; ML systems and RL |
Requirements:
- Application stating the preferred placement length
- Recruiter screen
- Technical assessment
- Research discussion
- Reference check
Source · alignment.anthropic.com/2026/anthropic-residency
Published 30 October 2026 · Checked against the organizer's page 5 November 2026
DevGrants Daily, run by Albert Andrejev. How entries are made and checked.
Quick answers
Is visa sponsorship available?
Yes, for placements of 3 months or longer. Residents on 6 week placements are encouraged to work remotely from a country where Anthropic can hire.
Is there a separate application for the UK?
Yes. The UK posting is a remote role with optional use of the London office.