far.ai

FAR.AI is a non-profit AI research institute dedicated to ensuring that advanced artificial intelligence remains trustworthy, secure, and beneficial for everyone. The organization conducts technical research across robustness, deception, and red-teaming, while also carrying out pre-deployment testing for frontier developers like OpenAI and independent evaluations for governments such as the EU AI Office. Its efforts focus on closing the research-adoption gap by developing solutions to high-leverage technical safety challenges and scaling them to real-world deployment.

FAR.AI serves the broader AI ecosystem by hosting global events, operating FAR.Labs as a collaborative co-working space in Berkeley, and providing targeted grantmaking to academics and independent researchers. The institute's machine learning and technical work explores a portfolio of high-potential agendas, including capability control, adversarial attacks on language model safeguard pipelines, and mapping where honesty emerges in reinforcement learning with verifiable rewards.

The organization is actively building out its teams across various functions, hiring for roles such as Research Scientist, Head of Engineering for Red Team, Tech Lead Manager and Senior Software Engineer for GPU Cluster Infrastructure, IT & Security Manager, and Research Lead for Pre-training Safety. The open positions include both office-based roles in Berkeley and international or US-based remote options, encompassing senior, technical lead, and management positions.