jobs

Research Scientist, AGI Safety and Alignment, DeepMind

The Artificial General Intelligence (AGI) Safety and Alignment Team (ASAT) aims to reduce existential and catastrophic risk from AGI and eventually Artificial Superintelligence (ASI). We research novel techniques and work with the rest of G

Company

Google

Location

London, UK

Status

Open

Posted

2026-09-25T11:09:13+00:00

Google DeepMind's AGI Safety and Alignment Team hires research scientists (and software engineers) to reduce catastrophic and existential risk from AGI. The work covers new alignment methods, alignment stress testing, language model interpretability, agent control, and amplified oversight, with sub-teams on control, debate-style training, and evaluation. Duties include applying techniques to frontier models, building production control systems, and advising executives through the frontier safety framework. Minimum: bachelor's degree, three years in software development, ML engineering or ML research, and experience with research teams. On-site in London.