Resemble AI builds AI trust and safety technology aimed at stopping AI misuse. The company's work concentrates on deepfake detection, AI content watermarking, and analysis of AI intent and behaviour, addressing problems in AI safety and security, trust and safety, and media authenticity. Its teams operate as a global distributed unit.
Its technical scope spans AI detection research, world models for detection, AI agent behaviour detection, multimodal watermarking, deepfake detection, and intent recognition and intelligence modelling. Four products sit at the centre of this work. DETECT-World is described as the first world model for detection and is positioned at the frontier of AI content detection research. AI agent behaviour detection targets the identification of how AI agents act. PerTh provides multimodal watermarking for marking and identifying AI-generated content. An intelligence model is designed to understand the intent behind AI-generated material.
The company describes a collaborative and inclusive working environment that values diverse perspectives, continuous learning, and accountability. Compensation is competitive and includes equity options, alongside a benefits package covering well-being and professional development. Advancement opportunities are offered as the company grows.
