Safety Research Internship (Spring/Summer 2026)

📋 Description

• • Embark on a groundbreaking research journey as a Safety Research Intern at Cohere Inc., a leading force in scaling intelligence for humanity. This internship offers a unique opportunity to collaborate directly with Cohere’s esteemed Modelling Safety team, contributing to the implementation and advancement of novel research ideas that are directly grounded in real-world usage and production-level AI models.
• • Dive deep into critical areas of AI safety and reliability, including but not limited to fairness, overall model safety across diverse languages and cultural contexts, robustness, generalization capabilities, interpretability, and the specific safety challenges posed by agents with complex read/write actions and code generation (codegen) tasks. If your research interests lie in an area not explicitly listed, you are strongly encouraged to propose it, as innovative ideas are highly valued.
• • To maximize your impact and align with the team's interests, it is recommended to familiarize yourself with the team's existing publications. This will provide valuable insights into the current research landscape and potential areas of focus.
• • The primary objective of this internship is to produce impactful research, aiming for publication in a top-tier academic venue and contributing to the broader open science community. While not a mandatory requirement, there is a significant possibility that your internship project could influence Cohere's production systems and model development, offering a tangible impact on cutting-edge AI products.
• • Your specific project details and research topic will be collaboratively designed in close partnership with experienced researchers on the Modelling Safety team, ensuring a project that is both personally engaging and strategically aligned with Cohere's mission.
• • As a Research Intern, you will be at the forefront of machine learning research, actively involved in training and evaluating large language models that are already in production. This hands-on experience will provide unparalleled exposure to the challenges and intricacies of developing and deploying advanced AI systems.
• • Your research will focus on making AI models more understandable, secure, dependable, and inclusive, ultimately ensuring they are more beneficial for society. This involves rigorous investigation into the ethical and practical implications of AI.
• • A key output of your internship will be the dissemination of your research findings. This will be achieved through the creation of high-quality publications, the development of novel datasets, and the release of associated code, fostering transparency and collaboration within the AI research community.
• • You will actively contribute to research initiatives that have direct, practical applications in Cohere’s product development, bridging the gap between theoretical research and real-world implementation.
• • This internship is designed for individuals pursuing a PhD in Machine Learning, Natural Language Processing (NLP), Artificial Intelligence, or a closely related discipline, though exceptional candidates not currently in a PhD program will also be considered. The role requires full-time commitment for a duration of 4 to 6 months, and it offers the flexibility of remote work, with the option to work from Cohere's office locations if geographically convenient.
• • You will gain invaluable experience in conducting cutting-edge machine learning research, focusing on the critical aspects of AI safety and reliability. This role is ideal for those passionate about shaping the future of AI and contributing to its responsible development.