Research Scientist Intern – Human-Centric Generative AI

Summary At Apple, we push the boundary of what is possible to improve the lives of our customers. The Europe Vision Group (EVG) is focused on enhancing communication by advancing the state-of-the-art in computer vision and machine learning. We are shipping features such as Persona for Apple Vision Pro, Portrait Mode, ARKit, Animoji and FaceTime Eye Contact. These breakthroughs are enabled by state-of-the-art research combined with world-class design. We are looking for a research intern who is excited to shape the next generation of Apple products and contribute to the next wave of transformative computer vision experiences.

Description This position requires highly-motivated individuals who want to join us in the fast-moving field of large-scale models for human-centric 3D representations. You will be responsible for developing, implementing, evaluating, and improving computer vision and GenAI algorithms that model photorealistic humans in complex and diverse scenarios.

Potential internship areas: Human-centric multi-modal generative models Human-centric video diffusion models Human behavior modeling and simulation Human-Object Interaction Human-Computer Interaction 3D reconstruction of humans Face and body tracking

Minimum qualifications • At the end of the internship, you must return to university to continue your education, or the internship must be the last requirement for you to graduate. • Strong understanding of modern machine learning techniques (e.g., generative, multi-modal, and foundational models; video understanding; and self-supervised learning). • Fluency in Python and Machine Learning frameworks (e.g. PyTorch) is required. • Excellent communication and collaboration skills. • Good problem-solving and analytical thinking abilities. • Working towards a PhD degree in Computer Science or a related field. PhD students are preferred for research-focused internships.

Preferred qualifications • Strong understanding of modern approaches for 3D reconstruction and rendering (e.g., Gaussian Splatting). • Hands-on experience with image or video diffusion models, 3D generative models, or vision foundation models. • Hands-on experience with human modeling algorithms (e.g., skeletal tracking, body shape and pose models, hair and clothes modeling). • Publication record at top conferences such as CVPR, ICCV, ECCV, NeurIPS, ICLR, or SIGGRAPH is a plus. • Industry experience is a plus. • Availability for 6 months minimum is preferred.