Member of Technical Staff, Safety for Agents

Who are we? Cohere is the leading security-first enterprise AI company. We build cutting-edge foundation AI models and end-to-end products that are designed to solve real-world business problems. We’re training and deploying frontier models for enterprises who are building AI systems. We believe that our work is instrumental to the widespread adoption of AI and we are looking for folks that want to be part of that. We obsess over what we build. Each one of us is responsible for contributing to increasing the capabilities of our models and the value they drive for our customers. Cohere is a team of researchers, engineers, designers, and more, who are all passionate about their craft. We are a global technology company co-headquartered in Toronto and San Francisco, with key offices in London, New York City, Montreal, Seoul, Germany and Paris. Join us! Why this role? As a Member of Technical Staff in the Safety for Agents team, you will make a meaningful impact on the development of better, fairer, more trustworthy, and more secure Large Language Models (LLMs). Your primary focus will be on data generation, post-training algorithms, and evaluation methods to ensure Safety in the next generation of models that can access external resources and take actions in the world. You will work closely with other cross-functional machine learning teams and data annotation teams, and will also collaborate with product and policy teams. This role combines expertise in machine learning, ethical and responsible AI, experimental design, and data generation and management. It will require curiosity to tackle totally new scientific problems, engineering skills to implement the pieces we need to test solutions to these, and a desire to dive into messy data and results. You will be on a small team with a lot of autonomy and decision-making power, responsible for making the next generation of LLMs better for society as a whole. Please Note: The existing team work in offices in London, Edinburgh, Paris, Toronto, and New York, but we also embrace being remote-friendly! For this role you need to have ~50% working day overlap with UK/EU timezone (e.g. US East is fine) but there are otherwise no restrictions on where you can be located for this role. You may be a good fit if you have: - Strong statistical skills and experience evaluating scientific experiments related to data collection and model performance. - Extremely strong software engineering skills. - Strong expertise in designing and conducting data collection tasks, including working with human annotators. - Experience analyzing datasets with respect to their quality, biases, and suitability for training ML models. - Hands-on experience training large language models (LLMs) on distributed training infrastructures. - Familiarity with evaluating and improving the generalizability and robustness of ML systems. - Proficiency in programming languages such as Python and ML frameworks (e.g., PyTorch, TensorFlow, JAX). - Excellent communication skills to collaborate effectively with cross-functional teams and present findings. - One or more papers at top-tier venues (such as NeurIPS, ICML, ICLR, AIStats, MLSys, JMLR, AAAI, Nature, COLING, ACL, EMNLP). HOW AND WHERE WE WORK: - Cohere is remote-friendly. We have offices in Toronto, San Francisco, New York City, London, Paris, Montreal, and more coming soon. - For those in the office: a daily lunch program, plenty of snacks, and regular community and social events. - For those not near an office: a co-working benefit so you can work alongside others in your city. If any of the above doesn’t line up exactly with your experience, we still encourage you to apply. We strive to create an inclusive work environment for all; we welcome applicants from all backgrounds and are committed to providing equal opportunities. Should you require any accommodations during the recruitment process, please submit an Accommodations Request Form https://docs.google.com/forms/d/12