Develop and implement computer vision systems and algorithms to analyse images and videos, enabling applications such as object detection, image recognition, tracking and automated visual inspection.
-Support and advise on research agendas on learning from human feedback over multimodal and temporal signals (e.g., video, audio, human interaction data). -Design and run experiments on subjective and temporal annotation—modelling inter-annotator variability, label uncertainty, and preference dynamics over time. -Develop novel methods at the intersection of preference learning, affective computing, and multimodal foundation models, with a focus on computer vision applications. -Collaborate with engineering teams to translate research prototypes into production-ready models and evaluation pipelines. -Establish rigorous evaluation methodologies and benchmarks for human-feedback quality, model alignment, and robustness. -Publish findings in top-tier venues (e.g., NeurIPS, ICML, ICLR, CHI) and represent the company in the research community. -Partner with engineering teams to translate research insights into product direction and real-world impact.
If multiple Education and Language Profiles are defined, please note that you must fit at least one of them, but not necessarily all.
