MATS is an educational seminar and independent research program that aims to grow the field of AI safety research.
Berkeley, California
MATS is an educational seminar and independent research program that aims to provide talented scholars with talks, workshops, and research mentorship in the field of AI safety, and connect them with the Berkeley AI safety research community. Past mentors include researchers from OpenAI, Anthropic, Google DeepMind, MIRI, ARC, CHAI, CAIS, Centre on Long-Term Risk, and more. MATS alumni have joined Anthropic, MIRI, ARC Evals, and CHAI, and founded several AI safety organisations, including Apollo Research, Leap Labs, and a new MIRI division.
Hybrid
Build and operate MATS's corporate IT security function, including endpoint protection, identity and access management, email security, compliance, vendor reviews, and incident response. Establish secure onboarding and offboarding for staff and fellows while enabling research teams to work efficiently and meet external partners' security requirements.
Hybrid
Design and implement security for research infrastructure, including Linux and SLURM clusters, multi-cloud environments, CI/CD pipelines, automated agents, and research data and model artefacts. Establish secure access, supply-chain controls, monitoring, incident response, and compliance practices while enabling researchers and external lab partnerships to work safely.