AI safety & policy
Fellowship
Deadline soon
Deadline
2026-11-15
Routes to impact
Direct high impact on an important cause
Testing your fit for a certain career path
Description
A two-week, part-time AI alignment research sprint: you propose a testable hypothesis on one of six research questions, and an engineering team runs the experiments.
- $1,000 researcher stipend per two-week sprint, for about 2 hours per week of your time
- 60+ engineering hours per sprint, plus cloud compute and frontier model API access
- You get working code, documented results, structured analysis and a final report
- Questions cover suppressed model hypotheses, escalation-risk signals, multi-agent failures and agent loyalty
- Promising sprints may extend an additional 1–6 months
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.
