Embark on a 12 to 24-month fixed-term contract as a Post-doctoral Researcher at DeepMind, contributing to the forefront of Artificial Intelligence. This role offers a unique opportunity to tackle complex AI challenges, collaborating with leading scientists and engineers.

The focus will be on human-computer interaction (HCI) research, exploring how individuals engage with, supervise, and provide feedback to advanced autonomous AI agents. You will be instrumental in designing and developing innovative interfaces and interaction techniques for effective human-agent collaboration, followed by rigorous user study evaluations.

The insights gained will directly influence how AI agents learn from human input, ensuring their safety and alignment with human intentions. This work holds the potential for significant positive impact on billions of users through future product applications.

Collaborate with research mentors to define impactful research projects and novel applications of machine learning, aligning with the team's objectives. Conduct groundbreaking research and publish high-caliber work. Design and implement large-scale experiments, developing reusable, production-ready code using frameworks like JAX. Deepen understanding of machine learning algorithms and apply research findings to drive design decisions. Execute user studies, including interviews, surveys, controlled experiments, and field deployments, to analyze human interaction with AI agents. Prototype novel interfaces and interaction methods facilitating human oversight and collaboration with autonomous agents. Develop human-centered evaluation methodologies and metrics for agent helpfulness, trustworthiness, safety, and alignment. Translate findings from human studies into enhanced AI agent learning mechanisms, such as reward modeling and human-in-the-loop reinforcement learning.

A PhD degree in a STEM field (Computer Science, Mathematics, Statistics) or equivalent practical experience is required. Proficiency in at least one general-purpose programming language such as Python, C/C++, or Java. Experience in training deep learning models using frameworks like JAX, TensorFlow, or PyTorch is essential. Demonstrated research experience in machine learning or AI techniques, evidenced by open-source projects, lab work, internships, or publications. Submission of at least one scientific publication to a top AI conference (e.g., NeurIPS, CVPR, ICCV, ICLR) is necessary.

Preferred qualifications include a Master’s degree in addition to the PhD, extensive research experience, collaborative skills across teams, and a strong publication record in premier conferences and journals.